8th day of python challenges 111-117

venv/lib/python3.6/site-packages/pandas/tests/reshape/merge/test_join.py

@@ -0,0 +1,862 @@
+import numpy as np
+from numpy.random import randn
+import pytest
+
+from pandas._libs import join as libjoin
+
+import pandas as pd
+from pandas import DataFrame, Index, MultiIndex, Series, concat, merge
+from pandas.tests.reshape.merge.test_merge import NGROUPS, N, get_test_data
+import pandas.util.testing as tm
+from pandas.util.testing import assert_frame_equal
+
+a_ = np.array
+
+
+class TestJoin:
+    def setup_method(self, method):
+        # aggregate multiple columns
+        self.df = DataFrame(
+            {
+                "key1": get_test_data(),
+                "key2": get_test_data(),
+                "data1": np.random.randn(N),
+                "data2": np.random.randn(N),
+            }
+        )
+
+        # exclude a couple keys for fun
+        self.df = self.df[self.df["key2"] > 1]
+
+        self.df2 = DataFrame(
+            {
+                "key1": get_test_data(n=N // 5),
+                "key2": get_test_data(ngroups=NGROUPS // 2, n=N // 5),
+                "value": np.random.randn(N // 5),
+            }
+        )
+
+        index, data = tm.getMixedTypeDict()
+        self.target = DataFrame(data, index=index)
+
+        # Join on string value
+        self.source = DataFrame(
+            {"MergedA": data["A"], "MergedD": data["D"]}, index=data["C"]
+        )
+
+    def test_cython_left_outer_join(self):
+        left = a_([0, 1, 2, 1, 2, 0, 0, 1, 2, 3, 3], dtype=np.int64)
+        right = a_([1, 1, 0, 4, 2, 2, 1], dtype=np.int64)
+        max_group = 5
+
+        ls, rs = libjoin.left_outer_join(left, right, max_group)
+
+        exp_ls = left.argsort(kind="mergesort")
+        exp_rs = right.argsort(kind="mergesort")
+
+        exp_li = a_([0, 1, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5, 6, 6, 7, 7, 8, 8, 9, 10])
+        exp_ri = a_([0, 0, 0, 1, 2, 3, 1, 2, 3, 1, 2, 3, 4, 5, 4, 5, 4, 5, -1, -1])
+
+        exp_ls = exp_ls.take(exp_li)
+        exp_ls[exp_li == -1] = -1
+
+        exp_rs = exp_rs.take(exp_ri)
+        exp_rs[exp_ri == -1] = -1
+
+        tm.assert_numpy_array_equal(ls, exp_ls, check_dtype=False)
+        tm.assert_numpy_array_equal(rs, exp_rs, check_dtype=False)
+
+    def test_cython_right_outer_join(self):
+        left = a_([0, 1, 2, 1, 2, 0, 0, 1, 2, 3, 3], dtype=np.int64)
+        right = a_([1, 1, 0, 4, 2, 2, 1], dtype=np.int64)
+        max_group = 5
+
+        rs, ls = libjoin.left_outer_join(right, left, max_group)
+
+        exp_ls = left.argsort(kind="mergesort")
+        exp_rs = right.argsort(kind="mergesort")
+
+        #            0        1        1        1
+        exp_li = a_(
+            [
+                0,
+                1,
+                2,
+                3,
+                4,
+                5,
+                3,
+                4,
+                5,
+                3,
+                4,
+                5,
+                #            2        2        4
+                6,
+                7,
+                8,
+                6,
+                7,
+                8,
+                -1,
+            ]
+        )
+        exp_ri = a_([0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5, 6])
+
+        exp_ls = exp_ls.take(exp_li)
+        exp_ls[exp_li == -1] = -1
+
+        exp_rs = exp_rs.take(exp_ri)
+        exp_rs[exp_ri == -1] = -1
+
+        tm.assert_numpy_array_equal(ls, exp_ls, check_dtype=False)
+        tm.assert_numpy_array_equal(rs, exp_rs, check_dtype=False)
+
+    def test_cython_inner_join(self):
+        left = a_([0, 1, 2, 1, 2, 0, 0, 1, 2, 3, 3], dtype=np.int64)
+        right = a_([1, 1, 0, 4, 2, 2, 1, 4], dtype=np.int64)
+        max_group = 5
+
+        ls, rs = libjoin.inner_join(left, right, max_group)
+
+        exp_ls = left.argsort(kind="mergesort")
+        exp_rs = right.argsort(kind="mergesort")
+
+        exp_li = a_([0, 1, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5, 6, 6, 7, 7, 8, 8])
+        exp_ri = a_([0, 0, 0, 1, 2, 3, 1, 2, 3, 1, 2, 3, 4, 5, 4, 5, 4, 5])
+
+        exp_ls = exp_ls.take(exp_li)
+        exp_ls[exp_li == -1] = -1
+
+        exp_rs = exp_rs.take(exp_ri)
+        exp_rs[exp_ri == -1] = -1
+
+        tm.assert_numpy_array_equal(ls, exp_ls, check_dtype=False)
+        tm.assert_numpy_array_equal(rs, exp_rs, check_dtype=False)
+
+    def test_left_outer_join(self):
+        joined_key2 = merge(self.df, self.df2, on="key2")
+        _check_join(self.df, self.df2, joined_key2, ["key2"], how="left")
+
+        joined_both = merge(self.df, self.df2)
+        _check_join(self.df, self.df2, joined_both, ["key1", "key2"], how="left")
+
+    def test_right_outer_join(self):
+        joined_key2 = merge(self.df, self.df2, on="key2", how="right")
+        _check_join(self.df, self.df2, joined_key2, ["key2"], how="right")
+
+        joined_both = merge(self.df, self.df2, how="right")
+        _check_join(self.df, self.df2, joined_both, ["key1", "key2"], how="right")
+
+    def test_full_outer_join(self):
+        joined_key2 = merge(self.df, self.df2, on="key2", how="outer")
+        _check_join(self.df, self.df2, joined_key2, ["key2"], how="outer")
+
+        joined_both = merge(self.df, self.df2, how="outer")
+        _check_join(self.df, self.df2, joined_both, ["key1", "key2"], how="outer")
+
+    def test_inner_join(self):
+        joined_key2 = merge(self.df, self.df2, on="key2", how="inner")
+        _check_join(self.df, self.df2, joined_key2, ["key2"], how="inner")
+
+        joined_both = merge(self.df, self.df2, how="inner")
+        _check_join(self.df, self.df2, joined_both, ["key1", "key2"], how="inner")
+
+    def test_handle_overlap(self):
+        joined = merge(self.df, self.df2, on="key2", suffixes=[".foo", ".bar"])
+
+        assert "key1.foo" in joined
+        assert "key1.bar" in joined
+
+    def test_handle_overlap_arbitrary_key(self):
+        joined = merge(
+            self.df,
+            self.df2,
+            left_on="key2",
+            right_on="key1",
+            suffixes=[".foo", ".bar"],
+        )
+        assert "key1.foo" in joined
+        assert "key2.bar" in joined
+
+    def test_join_on(self):
+        target = self.target
+        source = self.source
+
+        merged = target.join(source, on="C")
+        tm.assert_series_equal(merged["MergedA"], target["A"], check_names=False)
+        tm.assert_series_equal(merged["MergedD"], target["D"], check_names=False)
+
+        # join with duplicates (fix regression from DataFrame/Matrix merge)
+        df = DataFrame({"key": ["a", "a", "b", "b", "c"]})
+        df2 = DataFrame({"value": [0, 1, 2]}, index=["a", "b", "c"])
+        joined = df.join(df2, on="key")
+        expected = DataFrame(
+            {"key": ["a", "a", "b", "b", "c"], "value": [0, 0, 1, 1, 2]}
+        )
+        assert_frame_equal(joined, expected)
+
+        # Test when some are missing
+        df_a = DataFrame([[1], [2], [3]], index=["a", "b", "c"], columns=["one"])
+        df_b = DataFrame([["foo"], ["bar"]], index=[1, 2], columns=["two"])
+        df_c = DataFrame([[1], [2]], index=[1, 2], columns=["three"])
+        joined = df_a.join(df_b, on="one")
+        joined = joined.join(df_c, on="one")
+        assert np.isnan(joined["two"]["c"])
+        assert np.isnan(joined["three"]["c"])
+
+        # merge column not p resent
+        with pytest.raises(KeyError, match="^'E'$"):
+            target.join(source, on="E")
+
+        # overlap
+        source_copy = source.copy()
+        source_copy["A"] = 0
+        msg = (
+            "You are trying to merge on float64 and object columns. If"
+            " you wish to proceed you should use pd.concat"
+        )
+        with pytest.raises(ValueError, match=msg):
+            target.join(source_copy, on="A")
+
+    def test_join_on_fails_with_different_right_index(self):
+        df = DataFrame(
+            {"a": np.random.choice(["m", "f"], size=3), "b": np.random.randn(3)}
+        )
+        df2 = DataFrame(
+            {"a": np.random.choice(["m", "f"], size=10), "b": np.random.randn(10)},
+            index=tm.makeCustomIndex(10, 2),
+        )
+        msg = (
+            r"len\(left_on\) must equal the number of levels in the index" ' of "right"'
+        )
+        with pytest.raises(ValueError, match=msg):
+            merge(df, df2, left_on="a", right_index=True)
+
+    def test_join_on_fails_with_different_left_index(self):
+        df = DataFrame(
+            {"a": np.random.choice(["m", "f"], size=3), "b": np.random.randn(3)},
+            index=tm.makeCustomIndex(3, 2),
+        )
+        df2 = DataFrame(
+            {"a": np.random.choice(["m", "f"], size=10), "b": np.random.randn(10)}
+        )
+        msg = (
+            r"len\(right_on\) must equal the number of levels in the index" ' of "left"'
+        )
+        with pytest.raises(ValueError, match=msg):
+            merge(df, df2, right_on="b", left_index=True)
+
+    def test_join_on_fails_with_different_column_counts(self):
+        df = DataFrame(
+            {"a": np.random.choice(["m", "f"], size=3), "b": np.random.randn(3)}
+        )
+        df2 = DataFrame(
+            {"a": np.random.choice(["m", "f"], size=10), "b": np.random.randn(10)},
+            index=tm.makeCustomIndex(10, 2),
+        )
+        msg = r"len\(right_on\) must equal len\(left_on\)"
+        with pytest.raises(ValueError, match=msg):
+            merge(df, df2, right_on="a", left_on=["a", "b"])
+
+    @pytest.mark.parametrize("wrong_type", [2, "str", None, np.array([0, 1])])
+    def test_join_on_fails_with_wrong_object_type(self, wrong_type):
+        # GH12081 - original issue
+
+        # GH21220 - merging of Series and DataFrame is now allowed
+        # Edited test to remove the Series object from test parameters
+
+        df = DataFrame({"a": [1, 1]})
+        msg = "Can only merge Series or DataFrame objects, a {} was passed".format(
+            str(type(wrong_type))
+        )
+        with pytest.raises(TypeError, match=msg):
+            merge(wrong_type, df, left_on="a", right_on="a")
+        with pytest.raises(TypeError, match=msg):
+            merge(df, wrong_type, left_on="a", right_on="a")
+
+    def test_join_on_pass_vector(self):
+        expected = self.target.join(self.source, on="C")
+        del expected["C"]
+
+        join_col = self.target.pop("C")
+        result = self.target.join(self.source, on=join_col)
+        assert_frame_equal(result, expected)
+
+    def test_join_with_len0(self):
+        # nothing to merge
+        merged = self.target.join(self.source.reindex([]), on="C")
+        for col in self.source:
+            assert col in merged
+            assert merged[col].isna().all()
+
+        merged2 = self.target.join(self.source.reindex([]), on="C", how="inner")
+        tm.assert_index_equal(merged2.columns, merged.columns)
+        assert len(merged2) == 0
+
+    def test_join_on_inner(self):
+        df = DataFrame({"key": ["a", "a", "d", "b", "b", "c"]})
+        df2 = DataFrame({"value": [0, 1]}, index=["a", "b"])
+
+        joined = df.join(df2, on="key", how="inner")
+
+        expected = df.join(df2, on="key")
+        expected = expected[expected["value"].notna()]
+        tm.assert_series_equal(joined["key"], expected["key"], check_dtype=False)
+        tm.assert_series_equal(joined["value"], expected["value"], check_dtype=False)
+        tm.assert_index_equal(joined.index, expected.index)
+
+    def test_join_on_singlekey_list(self):
+        df = DataFrame({"key": ["a", "a", "b", "b", "c"]})
+        df2 = DataFrame({"value": [0, 1, 2]}, index=["a", "b", "c"])
+
+        # corner cases
+        joined = df.join(df2, on=["key"])
+        expected = df.join(df2, on="key")
+
+        assert_frame_equal(joined, expected)
+
+    def test_join_on_series(self):
+        result = self.target.join(self.source["MergedA"], on="C")
+        expected = self.target.join(self.source[["MergedA"]], on="C")
+        assert_frame_equal(result, expected)
+
+    def test_join_on_series_buglet(self):
+        # GH #638
+        df = DataFrame({"a": [1, 1]})
+        ds = Series([2], index=[1], name="b")
+        result = df.join(ds, on="a")
+        expected = DataFrame({"a": [1, 1], "b": [2, 2]}, index=df.index)
+        tm.assert_frame_equal(result, expected)
+
+    def test_join_index_mixed(self, join_type):
+        # no overlapping blocks
+        df1 = DataFrame(index=np.arange(10))
+        df1["bool"] = True
+        df1["string"] = "foo"
+
+        df2 = DataFrame(index=np.arange(5, 15))
+        df2["int"] = 1
+        df2["float"] = 1.0
+
+        joined = df1.join(df2, how=join_type)
+        expected = _join_by_hand(df1, df2, how=join_type)
+        assert_frame_equal(joined, expected)
+
+        joined = df2.join(df1, how=join_type)
+        expected = _join_by_hand(df2, df1, how=join_type)
+        assert_frame_equal(joined, expected)
+
+    def test_join_index_mixed_overlap(self):
+        df1 = DataFrame(
+            {"A": 1.0, "B": 2, "C": "foo", "D": True},
+            index=np.arange(10),
+            columns=["A", "B", "C", "D"],
+        )
+        assert df1["B"].dtype == np.int64
+        assert df1["D"].dtype == np.bool_
+
+        df2 = DataFrame(
+            {"A": 1.0, "B": 2, "C": "foo", "D": True},
+            index=np.arange(0, 10, 2),
+            columns=["A", "B", "C", "D"],
+        )
+
+        # overlap
+        joined = df1.join(df2, lsuffix="_one", rsuffix="_two")
+        expected_columns = [
+            "A_one",
+            "B_one",
+            "C_one",
+            "D_one",
+            "A_two",
+            "B_two",
+            "C_two",
+            "D_two",
+        ]
+        df1.columns = expected_columns[:4]
+        df2.columns = expected_columns[4:]
+        expected = _join_by_hand(df1, df2)
+        assert_frame_equal(joined, expected)
+
+    def test_join_empty_bug(self):
+        # generated an exception in 0.4.3
+        x = DataFrame()
+        x.join(DataFrame([3], index=[0], columns=["A"]), how="outer")
+
+    def test_join_unconsolidated(self):
+        # GH #331
+        a = DataFrame(randn(30, 2), columns=["a", "b"])
+        c = Series(randn(30))
+        a["c"] = c
+        d = DataFrame(randn(30, 1), columns=["q"])
+
+        # it works!
+        a.join(d)
+        d.join(a)
+
+    def test_join_multiindex(self):
+        index1 = MultiIndex.from_arrays(
+            [["a", "a", "a", "b", "b", "b"], [1, 2, 3, 1, 2, 3]],
+            names=["first", "second"],
+        )
+
+        index2 = MultiIndex.from_arrays(
+            [["b", "b", "b", "c", "c", "c"], [1, 2, 3, 1, 2, 3]],
+            names=["first", "second"],
+        )
+
+        df1 = DataFrame(data=np.random.randn(6), index=index1, columns=["var X"])
+        df2 = DataFrame(data=np.random.randn(6), index=index2, columns=["var Y"])
+
+        df1 = df1.sort_index(level=0)
+        df2 = df2.sort_index(level=0)
+
+        joined = df1.join(df2, how="outer")
+        ex_index = Index(index1.values).union(Index(index2.values))
+        expected = df1.reindex(ex_index).join(df2.reindex(ex_index))
+        expected.index.names = index1.names
+        assert_frame_equal(joined, expected)
+        assert joined.index.names == index1.names
+
+        df1 = df1.sort_index(level=1)
+        df2 = df2.sort_index(level=1)
+
+        joined = df1.join(df2, how="outer").sort_index(level=0)
+        ex_index = Index(index1.values).union(Index(index2.values))
+        expected = df1.reindex(ex_index).join(df2.reindex(ex_index))
+        expected.index.names = index1.names
+
+        assert_frame_equal(joined, expected)
+        assert joined.index.names == index1.names
+
+    def test_join_inner_multiindex(self):
+        key1 = ["bar", "bar", "bar", "foo", "foo", "baz", "baz", "qux", "qux", "snap"]
+        key2 = [
+            "two",
+            "one",
+            "three",
+            "one",
+            "two",
+            "one",
+            "two",
+            "two",
+            "three",
+            "one",
+        ]
+
+        data = np.random.randn(len(key1))
+        data = DataFrame({"key1": key1, "key2": key2, "data": data})
+
+        index = MultiIndex(
+            levels=[["foo", "bar", "baz", "qux"], ["one", "two", "three"]],
+            codes=[[0, 0, 0, 1, 1, 2, 2, 3, 3, 3], [0, 1, 2, 0, 1, 1, 2, 0, 1, 2]],
+            names=["first", "second"],
+        )
+        to_join = DataFrame(
+            np.random.randn(10, 3), index=index, columns=["j_one", "j_two", "j_three"]
+        )
+
+        joined = data.join(to_join, on=["key1", "key2"], how="inner")
+        expected = merge(
+            data,
+            to_join.reset_index(),
+            left_on=["key1", "key2"],
+            right_on=["first", "second"],
+            how="inner",
+            sort=False,
+        )
+
+        expected2 = merge(
+            to_join,
+            data,
+            right_on=["key1", "key2"],
+            left_index=True,
+            how="inner",
+            sort=False,
+        )
+        assert_frame_equal(joined, expected2.reindex_like(joined))
+
+        expected2 = merge(
+            to_join,
+            data,
+            right_on=["key1", "key2"],
+            left_index=True,
+            how="inner",
+            sort=False,
+        )
+
+        expected = expected.drop(["first", "second"], axis=1)
+        expected.index = joined.index
+
+        assert joined.index.is_monotonic
+        assert_frame_equal(joined, expected)
+
+        # _assert_same_contents(expected, expected2.loc[:, expected.columns])
+
+    def test_join_hierarchical_mixed(self):
+        # GH 2024
+        df = DataFrame([(1, 2, 3), (4, 5, 6)], columns=["a", "b", "c"])
+        new_df = df.groupby(["a"]).agg({"b": [np.mean, np.sum]})
+        other_df = DataFrame([(1, 2, 3), (7, 10, 6)], columns=["a", "b", "d"])
+        other_df.set_index("a", inplace=True)
+        # GH 9455, 12219
+        with tm.assert_produces_warning(UserWarning):
+            result = merge(new_df, other_df, left_index=True, right_index=True)
+        assert ("b", "mean") in result
+        assert "b" in result
+
+    def test_join_float64_float32(self):
+
+        a = DataFrame(randn(10, 2), columns=["a", "b"], dtype=np.float64)
+        b = DataFrame(randn(10, 1), columns=["c"], dtype=np.float32)
+        joined = a.join(b)
+        assert joined.dtypes["a"] == "float64"
+        assert joined.dtypes["b"] == "float64"
+        assert joined.dtypes["c"] == "float32"
+
+        a = np.random.randint(0, 5, 100).astype("int64")
+        b = np.random.random(100).astype("float64")
+        c = np.random.random(100).astype("float32")
+        df = DataFrame({"a": a, "b": b, "c": c})
+        xpdf = DataFrame({"a": a, "b": b, "c": c})
+        s = DataFrame(np.random.random(5).astype("float32"), columns=["md"])
+        rs = df.merge(s, left_on="a", right_index=True)
+        assert rs.dtypes["a"] == "int64"
+        assert rs.dtypes["b"] == "float64"
+        assert rs.dtypes["c"] == "float32"
+        assert rs.dtypes["md"] == "float32"
+
+        xp = xpdf.merge(s, left_on="a", right_index=True)
+        assert_frame_equal(rs, xp)
+
+    def test_join_many_non_unique_index(self):
+        df1 = DataFrame({"a": [1, 1], "b": [1, 1], "c": [10, 20]})
+        df2 = DataFrame({"a": [1, 1], "b": [1, 2], "d": [100, 200]})
+        df3 = DataFrame({"a": [1, 1], "b": [1, 2], "e": [1000, 2000]})
+        idf1 = df1.set_index(["a", "b"])
+        idf2 = df2.set_index(["a", "b"])
+        idf3 = df3.set_index(["a", "b"])
+
+        result = idf1.join([idf2, idf3], how="outer")
+
+        df_partially_merged = merge(df1, df2, on=["a", "b"], how="outer")
+        expected = merge(df_partially_merged, df3, on=["a", "b"], how="outer")
+
+        result = result.reset_index()
+        expected = expected[result.columns]
+        expected["a"] = expected.a.astype("int64")
+        expected["b"] = expected.b.astype("int64")
+        assert_frame_equal(result, expected)
+
+        df1 = DataFrame({"a": [1, 1, 1], "b": [1, 1, 1], "c": [10, 20, 30]})
+        df2 = DataFrame({"a": [1, 1, 1], "b": [1, 1, 2], "d": [100, 200, 300]})
+        df3 = DataFrame({"a": [1, 1, 1], "b": [1, 1, 2], "e": [1000, 2000, 3000]})
+        idf1 = df1.set_index(["a", "b"])
+        idf2 = df2.set_index(["a", "b"])
+        idf3 = df3.set_index(["a", "b"])
+        result = idf1.join([idf2, idf3], how="inner")
+
+        df_partially_merged = merge(df1, df2, on=["a", "b"], how="inner")
+        expected = merge(df_partially_merged, df3, on=["a", "b"], how="inner")
+
+        result = result.reset_index()
+
+        assert_frame_equal(result, expected.loc[:, result.columns])
+
+        # GH 11519
+        df = DataFrame(
+            {
+                "A": ["foo", "bar", "foo", "bar", "foo", "bar", "foo", "foo"],
+                "B": ["one", "one", "two", "three", "two", "two", "one", "three"],
+                "C": np.random.randn(8),
+                "D": np.random.randn(8),
+            }
+        )
+        s = Series(
+            np.repeat(np.arange(8), 2), index=np.repeat(np.arange(8), 2), name="TEST"
+        )
+        inner = df.join(s, how="inner")
+        outer = df.join(s, how="outer")
+        left = df.join(s, how="left")
+        right = df.join(s, how="right")
+        assert_frame_equal(inner, outer)
+        assert_frame_equal(inner, left)
+        assert_frame_equal(inner, right)
+
+    def test_join_sort(self):
+        left = DataFrame({"key": ["foo", "bar", "baz", "foo"], "value": [1, 2, 3, 4]})
+        right = DataFrame({"value2": ["a", "b", "c"]}, index=["bar", "baz", "foo"])
+
+        joined = left.join(right, on="key", sort=True)
+        expected = DataFrame(
+            {
+                "key": ["bar", "baz", "foo", "foo"],
+                "value": [2, 3, 1, 4],
+                "value2": ["a", "b", "c", "c"],
+            },
+            index=[1, 2, 0, 3],
+        )
+        assert_frame_equal(joined, expected)
+
+        # smoke test
+        joined = left.join(right, on="key", sort=False)
+        tm.assert_index_equal(joined.index, pd.Index(list(range(4))))
+
+    def test_join_mixed_non_unique_index(self):
+        # GH 12814, unorderable types in py3 with a non-unique index
+        df1 = DataFrame({"a": [1, 2, 3, 4]}, index=[1, 2, 3, "a"])
+        df2 = DataFrame({"b": [5, 6, 7, 8]}, index=[1, 3, 3, 4])
+        result = df1.join(df2)
+        expected = DataFrame(
+            {"a": [1, 2, 3, 3, 4], "b": [5, np.nan, 6, 7, np.nan]},
+            index=[1, 2, 3, 3, "a"],
+        )
+        tm.assert_frame_equal(result, expected)
+
+        df3 = DataFrame({"a": [1, 2, 3, 4]}, index=[1, 2, 2, "a"])
+        df4 = DataFrame({"b": [5, 6, 7, 8]}, index=[1, 2, 3, 4])
+        result = df3.join(df4)
+        expected = DataFrame(
+            {"a": [1, 2, 3, 4], "b": [5, 6, 6, np.nan]}, index=[1, 2, 2, "a"]
+        )
+        tm.assert_frame_equal(result, expected)
+
+    def test_join_non_unique_period_index(self):
+        # GH #16871
+        index = pd.period_range("2016-01-01", periods=16, freq="M")
+        df = DataFrame([i for i in range(len(index))], index=index, columns=["pnum"])
+        df2 = concat([df, df])
+        result = df.join(df2, how="inner", rsuffix="_df2")
+        expected = DataFrame(
+            np.tile(np.arange(16, dtype=np.int64).repeat(2).reshape(-1, 1), 2),
+            columns=["pnum", "pnum_df2"],
+            index=df2.sort_index().index,
+        )
+        tm.assert_frame_equal(result, expected)
+
+    def test_mixed_type_join_with_suffix(self):
+        # GH #916
+        df = DataFrame(np.random.randn(20, 6), columns=["a", "b", "c", "d", "e", "f"])
+        df.insert(0, "id", 0)
+        df.insert(5, "dt", "foo")
+
+        grouped = df.groupby("id")
+        mn = grouped.mean()
+        cn = grouped.count()
+
+        # it works!
+        mn.join(cn, rsuffix="_right")
+
+    def test_join_many(self):
+        df = DataFrame(np.random.randn(10, 6), columns=list("abcdef"))
+        df_list = [df[["a", "b"]], df[["c", "d"]], df[["e", "f"]]]
+
+        joined = df_list[0].join(df_list[1:])
+        tm.assert_frame_equal(joined, df)
+
+        df_list = [df[["a", "b"]][:-2], df[["c", "d"]][2:], df[["e", "f"]][1:9]]
+
+        def _check_diff_index(df_list, result, exp_index):
+            reindexed = [x.reindex(exp_index) for x in df_list]
+            expected = reindexed[0].join(reindexed[1:])
+            tm.assert_frame_equal(result, expected)
+
+        # different join types
+        joined = df_list[0].join(df_list[1:], how="outer")
+        _check_diff_index(df_list, joined, df.index)
+
+        joined = df_list[0].join(df_list[1:])
+        _check_diff_index(df_list, joined, df_list[0].index)
+
+        joined = df_list[0].join(df_list[1:], how="inner")
+        _check_diff_index(df_list, joined, df.index[2:8])
+
+        msg = "Joining multiple DataFrames only supported for joining on index"
+        with pytest.raises(ValueError, match=msg):
+            df_list[0].join(df_list[1:], on="a")
+
+    def test_join_many_mixed(self):
+        df = DataFrame(np.random.randn(8, 4), columns=["A", "B", "C", "D"])
+        df["key"] = ["foo", "bar"] * 4
+        df1 = df.loc[:, ["A", "B"]]
+        df2 = df.loc[:, ["C", "D"]]
+        df3 = df.loc[:, ["key"]]
+
+        result = df1.join([df2, df3])
+        assert_frame_equal(result, df)
+
+    def test_join_dups(self):
+
+        # joining dups
+        df = concat(
+            [
+                DataFrame(np.random.randn(10, 4), columns=["A", "A", "B", "B"]),
+                DataFrame(
+                    np.random.randint(0, 10, size=20).reshape(10, 2), columns=["A", "C"]
+                ),
+            ],
+            axis=1,
+        )
+
+        expected = concat([df, df], axis=1)
+        result = df.join(df, rsuffix="_2")
+        result.columns = expected.columns
+        assert_frame_equal(result, expected)
+
+        # GH 4975, invalid join on dups
+        w = DataFrame(np.random.randn(4, 2), columns=["x", "y"])
+        x = DataFrame(np.random.randn(4, 2), columns=["x", "y"])
+        y = DataFrame(np.random.randn(4, 2), columns=["x", "y"])
+        z = DataFrame(np.random.randn(4, 2), columns=["x", "y"])
+
+        dta = x.merge(y, left_index=True, right_index=True).merge(
+            z, left_index=True, right_index=True, how="outer"
+        )
+        dta = dta.merge(w, left_index=True, right_index=True)
+        expected = concat([x, y, z, w], axis=1)
+        expected.columns = ["x_x", "y_x", "x_y", "y_y", "x_x", "y_x", "x_y", "y_y"]
+        assert_frame_equal(dta, expected)
+
+    def test_join_multi_to_multi(self, join_type):
+        # GH 20475
+        leftindex = MultiIndex.from_product(
+            [list("abc"), list("xy"), [1, 2]], names=["abc", "xy", "num"]
+        )
+        left = DataFrame({"v1": range(12)}, index=leftindex)
+
+        rightindex = MultiIndex.from_product(
+            [list("abc"), list("xy")], names=["abc", "xy"]
+        )
+        right = DataFrame({"v2": [100 * i for i in range(1, 7)]}, index=rightindex)
+
+        result = left.join(right, on=["abc", "xy"], how=join_type)
+        expected = (
+            left.reset_index()
+            .merge(right.reset_index(), on=["abc", "xy"], how=join_type)
+            .set_index(["abc", "xy", "num"])
+        )
+        assert_frame_equal(expected, result)
+
+        msg = (
+            r"len\(left_on\) must equal the number of levels in the index" ' of "right"'
+        )
+        with pytest.raises(ValueError, match=msg):
+            left.join(right, on="xy", how=join_type)
+
+        with pytest.raises(ValueError, match=msg):
+            right.join(left, on=["abc", "xy"], how=join_type)
+
+    def test_join_on_tz_aware_datetimeindex(self):
+        # GH 23931, 26335
+        df1 = pd.DataFrame(
+            {
+                "date": pd.date_range(
+                    start="2018-01-01", periods=5, tz="America/Chicago"
+                ),
+                "vals": list("abcde"),
+            }
+        )
+
+        df2 = pd.DataFrame(
+            {
+                "date": pd.date_range(
+                    start="2018-01-03", periods=5, tz="America/Chicago"
+                ),
+                "vals_2": list("tuvwx"),
+            }
+        )
+        result = df1.join(df2.set_index("date"), on="date")
+        expected = df1.copy()
+        expected["vals_2"] = pd.Series([np.nan] * 2 + list("tuv"), dtype=object)
+        assert_frame_equal(result, expected)
+
+
+def _check_join(left, right, result, join_col, how="left", lsuffix="_x", rsuffix="_y"):
+
+    # some smoke tests
+    for c in join_col:
+        assert result[c].notna().all()
+
+    left_grouped = left.groupby(join_col)
+    right_grouped = right.groupby(join_col)
+
+    for group_key, group in result.groupby(join_col):
+        l_joined = _restrict_to_columns(group, left.columns, lsuffix)
+        r_joined = _restrict_to_columns(group, right.columns, rsuffix)
+
+        try:
+            lgroup = left_grouped.get_group(group_key)
+        except KeyError:
+            if how in ("left", "inner"):
+                raise AssertionError(
+                    "key {group_key!s} should not have been in the join".format(
+                        group_key=group_key
+                    )
+                )
+
+            _assert_all_na(l_joined, left.columns, join_col)
+        else:
+            _assert_same_contents(l_joined, lgroup)
+
+        try:
+            rgroup = right_grouped.get_group(group_key)
+        except KeyError:
+            if how in ("right", "inner"):
+                raise AssertionError(
+                    "key {group_key!s} should not have been in the join".format(
+                        group_key=group_key
+                    )
+                )
+
+            _assert_all_na(r_joined, right.columns, join_col)
+        else:
+            _assert_same_contents(r_joined, rgroup)
+
+
+def _restrict_to_columns(group, columns, suffix):
+    found = [
+        c for c in group.columns if c in columns or c.replace(suffix, "") in columns
+    ]
+
+    # filter
+    group = group.loc[:, found]
+
+    # get rid of suffixes, if any
+    group = group.rename(columns=lambda x: x.replace(suffix, ""))
+
+    # put in the right order...
+    group = group.loc[:, columns]
+
+    return group
+
+
+def _assert_same_contents(join_chunk, source):
+    NA_SENTINEL = -1234567  # drop_duplicates not so NA-friendly...
+
+    jvalues = join_chunk.fillna(NA_SENTINEL).drop_duplicates().values
+    svalues = source.fillna(NA_SENTINEL).drop_duplicates().values
+
+    rows = {tuple(row) for row in jvalues}
+    assert len(rows) == len(source)
+    assert all(tuple(row) in rows for row in svalues)
+
+
+def _assert_all_na(join_chunk, source_columns, join_col):
+    for c in source_columns:
+        if c in join_col:
+            continue
+        assert join_chunk[c].isna().all()
+
+
+def _join_by_hand(a, b, how="left"):
+    join_index = a.index.join(b.index, how=how)
+
+    a_re = a.reindex(join_index)
+    b_re = b.reindex(join_index)
+
+    result_columns = a.columns.append(b.columns)
+
+    for col, s in b_re.items():
+        a_re[col] = s
+    return a_re.reindex(columns=result_columns)

venv/lib/python3.6/site-packages/pandas/tests/reshape/merge/test_merge_index_as_string.py

@@ -0,0 +1,188 @@
+import numpy as np
+import pytest
+
+from pandas import DataFrame
+from pandas.util.testing import assert_frame_equal
+
+
+@pytest.fixture
+def df1():
+    return DataFrame(
+        dict(
+            outer=[1, 1, 1, 2, 2, 2, 2, 3, 3, 4, 4],
+            inner=[1, 2, 3, 1, 2, 3, 4, 1, 2, 1, 2],
+            v1=np.linspace(0, 1, 11),
+        )
+    )
+
+
+@pytest.fixture
+def df2():
+    return DataFrame(
+        dict(
+            outer=[1, 1, 1, 1, 1, 1, 2, 2, 3, 3, 3, 3],
+            inner=[1, 2, 2, 3, 3, 4, 2, 3, 1, 1, 2, 3],
+            v2=np.linspace(10, 11, 12),
+        )
+    )
+
+
+@pytest.fixture(params=[[], ["outer"], ["outer", "inner"]])
+def left_df(request, df1):
+    """ Construct left test DataFrame with specified levels
+    (any of 'outer', 'inner', and 'v1')"""
+    levels = request.param
+    if levels:
+        df1 = df1.set_index(levels)
+
+    return df1
+
+
+@pytest.fixture(params=[[], ["outer"], ["outer", "inner"]])
+def right_df(request, df2):
+    """ Construct right test DataFrame with specified levels
+    (any of 'outer', 'inner', and 'v2')"""
+    levels = request.param
+
+    if levels:
+        df2 = df2.set_index(levels)
+
+    return df2
+
+
+def compute_expected(df_left, df_right, on=None, left_on=None, right_on=None, how=None):
+    """
+    Compute the expected merge result for the test case.
+
+    This method computes the expected result of merging two DataFrames on
+    a combination of their columns and index levels. It does so by
+    explicitly dropping/resetting their named index levels, performing a
+    merge on their columns, and then finally restoring the appropriate
+    index in the result.
+
+    Parameters
+    ----------
+    df_left : DataFrame
+        The left DataFrame (may have zero or more named index levels)
+    df_right : DataFrame
+        The right DataFrame (may have zero or more named index levels)
+    on : list of str
+        The on parameter to the merge operation
+    left_on : list of str
+        The left_on parameter to the merge operation
+    right_on : list of str
+        The right_on parameter to the merge operation
+    how : str
+        The how parameter to the merge operation
+
+    Returns
+    -------
+    DataFrame
+        The expected merge result
+    """
+
+    # Handle on param if specified
+    if on is not None:
+        left_on, right_on = on, on
+
+    # Compute input named index levels
+    left_levels = [n for n in df_left.index.names if n is not None]
+    right_levels = [n for n in df_right.index.names if n is not None]
+
+    # Compute output named index levels
+    output_levels = [i for i in left_on if i in right_levels and i in left_levels]
+
+    # Drop index levels that aren't involved in the merge
+    drop_left = [n for n in left_levels if n not in left_on]
+    if drop_left:
+        df_left = df_left.reset_index(drop_left, drop=True)
+
+    drop_right = [n for n in right_levels if n not in right_on]
+    if drop_right:
+        df_right = df_right.reset_index(drop_right, drop=True)
+
+    # Convert remaining index levels to columns
+    reset_left = [n for n in left_levels if n in left_on]
+    if reset_left:
+        df_left = df_left.reset_index(level=reset_left)
+
+    reset_right = [n for n in right_levels if n in right_on]
+    if reset_right:
+        df_right = df_right.reset_index(level=reset_right)
+
+    # Perform merge
+    expected = df_left.merge(df_right, left_on=left_on, right_on=right_on, how=how)
+
+    # Restore index levels
+    if output_levels:
+        expected = expected.set_index(output_levels)
+
+    return expected
+
+
+@pytest.mark.parametrize(
+    "on,how",
+    [
+        (["outer"], "inner"),
+        (["inner"], "left"),
+        (["outer", "inner"], "right"),
+        (["inner", "outer"], "outer"),
+    ],
+)
+def test_merge_indexes_and_columns_on(left_df, right_df, on, how):
+
+    # Construct expected result
+    expected = compute_expected(left_df, right_df, on=on, how=how)
+
+    # Perform merge
+    result = left_df.merge(right_df, on=on, how=how)
+    assert_frame_equal(result, expected, check_like=True)
+
+
+@pytest.mark.parametrize(
+    "left_on,right_on,how",
+    [
+        (["outer"], ["outer"], "inner"),
+        (["inner"], ["inner"], "right"),
+        (["outer", "inner"], ["outer", "inner"], "left"),
+        (["inner", "outer"], ["inner", "outer"], "outer"),
+    ],
+)
+def test_merge_indexes_and_columns_lefton_righton(
+    left_df, right_df, left_on, right_on, how
+):
+
+    # Construct expected result
+    expected = compute_expected(
+        left_df, right_df, left_on=left_on, right_on=right_on, how=how
+    )
+
+    # Perform merge
+    result = left_df.merge(right_df, left_on=left_on, right_on=right_on, how=how)
+    assert_frame_equal(result, expected, check_like=True)
+
+
+@pytest.mark.parametrize("left_index", ["inner", ["inner", "outer"]])
+def test_join_indexes_and_columns_on(df1, df2, left_index, join_type):
+
+    # Construct left_df
+    left_df = df1.set_index(left_index)
+
+    # Construct right_df
+    right_df = df2.set_index(["outer", "inner"])
+
+    # Result
+    expected = (
+        left_df.reset_index()
+        .join(
+            right_df, on=["outer", "inner"], how=join_type, lsuffix="_x", rsuffix="_y"
+        )
+        .set_index(left_index)
+    )
+
+    # Perform join
+    result = left_df.join(
+        right_df, on=["outer", "inner"], how=join_type, lsuffix="_x", rsuffix="_y"
+    )
+
+    assert_frame_equal(result, expected, check_like=True)

venv/lib/python3.6/site-packages/pandas/tests/reshape/merge/test_merge_ordered.py

@@ -0,0 +1,117 @@
+from numpy import nan
+import pytest
+
+import pandas as pd
+from pandas import DataFrame, merge_ordered
+from pandas.util.testing import assert_frame_equal
+
+
+class TestMergeOrdered:
+    def setup_method(self, method):
+        self.left = DataFrame({"key": ["a", "c", "e"], "lvalue": [1, 2.0, 3]})
+
+        self.right = DataFrame({"key": ["b", "c", "d", "f"], "rvalue": [1, 2, 3.0, 4]})
+
+    def test_basic(self):
+        result = merge_ordered(self.left, self.right, on="key")
+        expected = DataFrame(
+            {
+                "key": ["a", "b", "c", "d", "e", "f"],
+                "lvalue": [1, nan, 2, nan, 3, nan],
+                "rvalue": [nan, 1, 2, 3, nan, 4],
+            }
+        )
+
+        assert_frame_equal(result, expected)
+
+    def test_ffill(self):
+        result = merge_ordered(self.left, self.right, on="key", fill_method="ffill")
+        expected = DataFrame(
+            {
+                "key": ["a", "b", "c", "d", "e", "f"],
+                "lvalue": [1.0, 1, 2, 2, 3, 3.0],
+                "rvalue": [nan, 1, 2, 3, 3, 4],
+            }
+        )
+        assert_frame_equal(result, expected)
+
+    def test_multigroup(self):
+        left = pd.concat([self.left, self.left], ignore_index=True)
+
+        left["group"] = ["a"] * 3 + ["b"] * 3
+
+        result = merge_ordered(
+            left, self.right, on="key", left_by="group", fill_method="ffill"
+        )
+        expected = DataFrame(
+            {
+                "key": ["a", "b", "c", "d", "e", "f"] * 2,
+                "lvalue": [1.0, 1, 2, 2, 3, 3.0] * 2,
+                "rvalue": [nan, 1, 2, 3, 3, 4] * 2,
+            }
+        )
+        expected["group"] = ["a"] * 6 + ["b"] * 6
+
+        assert_frame_equal(result, expected.loc[:, result.columns])
+
+        result2 = merge_ordered(
+            self.right, left, on="key", right_by="group", fill_method="ffill"
+        )
+        assert_frame_equal(result, result2.loc[:, result.columns])
+
+        result = merge_ordered(left, self.right, on="key", left_by="group")
+        assert result["group"].notna().all()
+
+    def test_merge_type(self):
+        class NotADataFrame(DataFrame):
+            @property
+            def _constructor(self):
+                return NotADataFrame
+
+        nad = NotADataFrame(self.left)
+        result = nad.merge(self.right, on="key")
+
+        assert isinstance(result, NotADataFrame)
+
+    def test_empty_sequence_concat(self):
+        # GH 9157
+        empty_pat = "[Nn]o objects"
+        none_pat = "objects.*None"
+        test_cases = [
+            ((), empty_pat),
+            ([], empty_pat),
+            ({}, empty_pat),
+            ([None], none_pat),
+            ([None, None], none_pat),
+        ]
+        for df_seq, pattern in test_cases:
+            with pytest.raises(ValueError, match=pattern):
+                pd.concat(df_seq)
+
+        pd.concat([pd.DataFrame()])
+        pd.concat([None, pd.DataFrame()])
+        pd.concat([pd.DataFrame(), None])
+
+    def test_doc_example(self):
+        left = DataFrame(
+            {
+                "group": list("aaabbb"),
+                "key": ["a", "c", "e", "a", "c", "e"],
+                "lvalue": [1, 2, 3] * 2,
+            }
+        )
+
+        right = DataFrame({"key": ["b", "c", "d"], "rvalue": [1, 2, 3]})
+
+        result = merge_ordered(left, right, fill_method="ffill", left_by="group")
+
+        expected = DataFrame(
+            {
+                "group": list("aaaaabbbbb"),
+                "key": ["a", "b", "c", "d", "e"] * 2,
+                "lvalue": [1, 1, 2, 2, 3] * 2,
+                "rvalue": [nan, 1, 2, 3, 3] * 2,
+            }
+        )
+
+        assert_frame_equal(result, expected)

venv/lib/python3.6/site-packages/pandas/tests/reshape/merge/test_multi.py

@@ -0,0 +1,810 @@
+from collections import OrderedDict
+
+import numpy as np
+from numpy import nan
+from numpy.random import randn
+import pytest
+
+import pandas as pd
+from pandas import DataFrame, Index, MultiIndex, Series
+from pandas.core.reshape.concat import concat
+from pandas.core.reshape.merge import merge
+import pandas.util.testing as tm
+
+
+@pytest.fixture
+def left():
+    """left dataframe (not multi-indexed) for multi-index join tests"""
+    # a little relevant example with NAs
+    key1 = ["bar", "bar", "bar", "foo", "foo", "baz", "baz", "qux", "qux", "snap"]
+    key2 = ["two", "one", "three", "one", "two", "one", "two", "two", "three", "one"]
+
+    data = np.random.randn(len(key1))
+    return DataFrame({"key1": key1, "key2": key2, "data": data})
+
+
+@pytest.fixture
+def right():
+    """right dataframe (multi-indexed) for multi-index join tests"""
+    index = MultiIndex(
+        levels=[["foo", "bar", "baz", "qux"], ["one", "two", "three"]],
+        codes=[[0, 0, 0, 1, 1, 2, 2, 3, 3, 3], [0, 1, 2, 0, 1, 1, 2, 0, 1, 2]],
+        names=["key1", "key2"],
+    )
+
+    return DataFrame(
+        np.random.randn(10, 3), index=index, columns=["j_one", "j_two", "j_three"]
+    )
+
+
+@pytest.fixture
+def left_multi():
+    return DataFrame(
+        dict(
+            Origin=["A", "A", "B", "B", "C"],
+            Destination=["A", "B", "A", "C", "A"],
+            Period=["AM", "AM", "IP", "AM", "OP"],
+            TripPurp=["hbw", "nhb", "hbo", "nhb", "hbw"],
+            Trips=[1987, 3647, 2470, 4296, 4444],
+        ),
+        columns=["Origin", "Destination", "Period", "TripPurp", "Trips"],
+    ).set_index(["Origin", "Destination", "Period", "TripPurp"])
+
+
+@pytest.fixture
+def right_multi():
+    return DataFrame(
+        dict(
+            Origin=["A", "A", "B", "B", "C", "C", "E"],
+            Destination=["A", "B", "A", "B", "A", "B", "F"],
+            Period=["AM", "AM", "IP", "AM", "OP", "IP", "AM"],
+            LinkType=["a", "b", "c", "b", "a", "b", "a"],
+            Distance=[100, 80, 90, 80, 75, 35, 55],
+        ),
+        columns=["Origin", "Destination", "Period", "LinkType", "Distance"],
+    ).set_index(["Origin", "Destination", "Period", "LinkType"])
+
+
+@pytest.fixture
+def on_cols_multi():
+    return ["Origin", "Destination", "Period"]
+
+
+@pytest.fixture
+def idx_cols_multi():
+    return ["Origin", "Destination", "Period", "TripPurp", "LinkType"]
+
+
+class TestMergeMulti:
+    def setup_method(self):
+        self.index = MultiIndex(
+            levels=[["foo", "bar", "baz", "qux"], ["one", "two", "three"]],
+            codes=[[0, 0, 0, 1, 1, 2, 2, 3, 3, 3], [0, 1, 2, 0, 1, 1, 2, 0, 1, 2]],
+            names=["first", "second"],
+        )
+        self.to_join = DataFrame(
+            np.random.randn(10, 3),
+            index=self.index,
+            columns=["j_one", "j_two", "j_three"],
+        )
+
+        # a little relevant example with NAs
+        key1 = ["bar", "bar", "bar", "foo", "foo", "baz", "baz", "qux", "qux", "snap"]
+        key2 = [
+            "two",
+            "one",
+            "three",
+            "one",
+            "two",
+            "one",
+            "two",
+            "two",
+            "three",
+            "one",
+        ]
+
+        data = np.random.randn(len(key1))
+        self.data = DataFrame({"key1": key1, "key2": key2, "data": data})
+
+    def test_merge_on_multikey(self, left, right, join_type):
+        on_cols = ["key1", "key2"]
+        result = left.join(right, on=on_cols, how=join_type).reset_index(drop=True)
+
+        expected = pd.merge(left, right.reset_index(), on=on_cols, how=join_type)
+
+        tm.assert_frame_equal(result, expected)
+
+        result = left.join(right, on=on_cols, how=join_type, sort=True).reset_index(
+            drop=True
+        )
+
+        expected = pd.merge(
+            left, right.reset_index(), on=on_cols, how=join_type, sort=True
+        )
+
+        tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("sort", [False, True])
+    def test_left_join_multi_index(self, left, right, sort):
+        icols = ["1st", "2nd", "3rd"]
+
+        def bind_cols(df):
+            iord = lambda a: 0 if a != a else ord(a)
+            f = lambda ts: ts.map(iord) - ord("a")
+            return f(df["1st"]) + f(df["3rd"]) * 1e2 + df["2nd"].fillna(0) * 1e4
+
+        def run_asserts(left, right, sort):
+            res = left.join(right, on=icols, how="left", sort=sort)
+
+            assert len(left) < len(res) + 1
+            assert not res["4th"].isna().any()
+            assert not res["5th"].isna().any()
+
+            tm.assert_series_equal(res["4th"], -res["5th"], check_names=False)
+            result = bind_cols(res.iloc[:, :-2])
+            tm.assert_series_equal(res["4th"], result, check_names=False)
+            assert result.name is None
+
+            if sort:
+                tm.assert_frame_equal(res, res.sort_values(icols, kind="mergesort"))
+
+            out = merge(left, right.reset_index(), on=icols, sort=sort, how="left")
+
+            res.index = np.arange(len(res))
+            tm.assert_frame_equal(out, res)
+
+        lc = list(map(chr, np.arange(ord("a"), ord("z") + 1)))
+        left = DataFrame(np.random.choice(lc, (5000, 2)), columns=["1st", "3rd"])
+        left.insert(1, "2nd", np.random.randint(0, 1000, len(left)))
+
+        i = np.random.permutation(len(left))
+        right = left.iloc[i].copy()
+
+        left["4th"] = bind_cols(left)
+        right["5th"] = -bind_cols(right)
+        right.set_index(icols, inplace=True)
+
+        run_asserts(left, right, sort)
+
+        # inject some nulls
+        left.loc[1::23, "1st"] = np.nan
+        left.loc[2::37, "2nd"] = np.nan
+        left.loc[3::43, "3rd"] = np.nan
+        left["4th"] = bind_cols(left)
+
+        i = np.random.permutation(len(left))
+        right = left.iloc[i, :-1]
+        right["5th"] = -bind_cols(right)
+        right.set_index(icols, inplace=True)
+
+        run_asserts(left, right, sort)
+
+    @pytest.mark.parametrize("sort", [False, True])
+    def test_merge_right_vs_left(self, left, right, sort):
+        # compare left vs right merge with multikey
+        on_cols = ["key1", "key2"]
+        merged_left_right = left.merge(
+            right, left_on=on_cols, right_index=True, how="left", sort=sort
+        )
+
+        merge_right_left = right.merge(
+            left, right_on=on_cols, left_index=True, how="right", sort=sort
+        )
+
+        # Reorder columns
+        merge_right_left = merge_right_left[merged_left_right.columns]
+
+        tm.assert_frame_equal(merged_left_right, merge_right_left)
+
+    def test_compress_group_combinations(self):
+
+        # ~ 40000000 possible unique groups
+        key1 = tm.rands_array(10, 10000)
+        key1 = np.tile(key1, 2)
+        key2 = key1[::-1]
+
+        df = DataFrame({"key1": key1, "key2": key2, "value1": np.random.randn(20000)})
+
+        df2 = DataFrame(
+            {"key1": key1[::2], "key2": key2[::2], "value2": np.random.randn(10000)}
+        )
+
+        # just to hit the label compression code path
+        merge(df, df2, how="outer")
+
+    def test_left_join_index_preserve_order(self):
+
+        on_cols = ["k1", "k2"]
+        left = DataFrame(
+            {
+                "k1": [0, 1, 2] * 8,
+                "k2": ["foo", "bar"] * 12,
+                "v": np.array(np.arange(24), dtype=np.int64),
+            }
+        )
+
+        index = MultiIndex.from_tuples([(2, "bar"), (1, "foo")])
+        right = DataFrame({"v2": [5, 7]}, index=index)
+
+        result = left.join(right, on=on_cols)
+
+        expected = left.copy()
+        expected["v2"] = np.nan
+        expected.loc[(expected.k1 == 2) & (expected.k2 == "bar"), "v2"] = 5
+        expected.loc[(expected.k1 == 1) & (expected.k2 == "foo"), "v2"] = 7
+
+        tm.assert_frame_equal(result, expected)
+
+        result.sort_values(on_cols, kind="mergesort", inplace=True)
+        expected = left.join(right, on=on_cols, sort=True)
+
+        tm.assert_frame_equal(result, expected)
+
+        # test join with multi dtypes blocks
+        left = DataFrame(
+            {
+                "k1": [0, 1, 2] * 8,
+                "k2": ["foo", "bar"] * 12,
+                "k3": np.array([0, 1, 2] * 8, dtype=np.float32),
+                "v": np.array(np.arange(24), dtype=np.int32),
+            }
+        )
+
+        index = MultiIndex.from_tuples([(2, "bar"), (1, "foo")])
+        right = DataFrame({"v2": [5, 7]}, index=index)
+
+        result = left.join(right, on=on_cols)
+
+        expected = left.copy()
+        expected["v2"] = np.nan
+        expected.loc[(expected.k1 == 2) & (expected.k2 == "bar"), "v2"] = 5
+        expected.loc[(expected.k1 == 1) & (expected.k2 == "foo"), "v2"] = 7
+
+        tm.assert_frame_equal(result, expected)
+
+        result = result.sort_values(on_cols, kind="mergesort")
+        expected = left.join(right, on=on_cols, sort=True)
+
+        tm.assert_frame_equal(result, expected)
+
+    def test_left_join_index_multi_match_multiindex(self):
+        left = DataFrame(
+            [
+                ["X", "Y", "C", "a"],
+                ["W", "Y", "C", "e"],
+                ["V", "Q", "A", "h"],
+                ["V", "R", "D", "i"],
+                ["X", "Y", "D", "b"],
+                ["X", "Y", "A", "c"],
+                ["W", "Q", "B", "f"],
+                ["W", "R", "C", "g"],
+                ["V", "Y", "C", "j"],
+                ["X", "Y", "B", "d"],
+            ],
+            columns=["cola", "colb", "colc", "tag"],
+            index=[3, 2, 0, 1, 7, 6, 4, 5, 9, 8],
+        )
+
+        right = DataFrame(
+            [
+                ["W", "R", "C", 0],
+                ["W", "Q", "B", 3],
+                ["W", "Q", "B", 8],
+                ["X", "Y", "A", 1],
+                ["X", "Y", "A", 4],
+                ["X", "Y", "B", 5],
+                ["X", "Y", "C", 6],
+                ["X", "Y", "C", 9],
+                ["X", "Q", "C", -6],
+                ["X", "R", "C", -9],
+                ["V", "Y", "C", 7],
+                ["V", "R", "D", 2],
+                ["V", "R", "D", -1],
+                ["V", "Q", "A", -3],
+            ],
+            columns=["col1", "col2", "col3", "val"],
+        ).set_index(["col1", "col2", "col3"])
+
+        result = left.join(right, on=["cola", "colb", "colc"], how="left")
+
+        expected = DataFrame(
+            [
+                ["X", "Y", "C", "a", 6],
+                ["X", "Y", "C", "a", 9],
+                ["W", "Y", "C", "e", nan],
+                ["V", "Q", "A", "h", -3],
+                ["V", "R", "D", "i", 2],
+                ["V", "R", "D", "i", -1],
+                ["X", "Y", "D", "b", nan],
+                ["X", "Y", "A", "c", 1],
+                ["X", "Y", "A", "c", 4],
+                ["W", "Q", "B", "f", 3],
+                ["W", "Q", "B", "f", 8],
+                ["W", "R", "C", "g", 0],
+                ["V", "Y", "C", "j", 7],
+                ["X", "Y", "B", "d", 5],
+            ],
+            columns=["cola", "colb", "colc", "tag", "val"],
+            index=[3, 3, 2, 0, 1, 1, 7, 6, 6, 4, 4, 5, 9, 8],
+        )
+
+        tm.assert_frame_equal(result, expected)
+
+        result = left.join(right, on=["cola", "colb", "colc"], how="left", sort=True)
+
+        expected = expected.sort_values(["cola", "colb", "colc"], kind="mergesort")
+
+        tm.assert_frame_equal(result, expected)
+
+    def test_left_join_index_multi_match(self):
+        left = DataFrame(
+            [["c", 0], ["b", 1], ["a", 2], ["b", 3]],
+            columns=["tag", "val"],
+            index=[2, 0, 1, 3],
+        )
+
+        right = DataFrame(
+            [
+                ["a", "v"],
+                ["c", "w"],
+                ["c", "x"],
+                ["d", "y"],
+                ["a", "z"],
+                ["c", "r"],
+                ["e", "q"],
+                ["c", "s"],
+            ],
+            columns=["tag", "char"],
+        ).set_index("tag")
+
+        result = left.join(right, on="tag", how="left")
+
+        expected = DataFrame(
+            [
+                ["c", 0, "w"],
+                ["c", 0, "x"],
+                ["c", 0, "r"],
+                ["c", 0, "s"],
+                ["b", 1, nan],
+                ["a", 2, "v"],
+                ["a", 2, "z"],
+                ["b", 3, nan],
+            ],
+            columns=["tag", "val", "char"],
+            index=[2, 2, 2, 2, 0, 1, 1, 3],
+        )
+
+        tm.assert_frame_equal(result, expected)
+
+        result = left.join(right, on="tag", how="left", sort=True)
+        expected2 = expected.sort_values("tag", kind="mergesort")
+
+        tm.assert_frame_equal(result, expected2)
+
+        # GH7331 - maintain left frame order in left merge
+        result = merge(left, right.reset_index(), how="left", on="tag")
+        expected.index = np.arange(len(expected))
+        tm.assert_frame_equal(result, expected)
+
+    def test_left_merge_na_buglet(self):
+        left = DataFrame(
+            {
+                "id": list("abcde"),
+                "v1": randn(5),
+                "v2": randn(5),
+                "dummy": list("abcde"),
+                "v3": randn(5),
+            },
+            columns=["id", "v1", "v2", "dummy", "v3"],
+        )
+        right = DataFrame(
+            {
+                "id": ["a", "b", np.nan, np.nan, np.nan],
+                "sv3": [1.234, 5.678, np.nan, np.nan, np.nan],
+            }
+        )
+
+        result = merge(left, right, on="id", how="left")
+
+        rdf = right.drop(["id"], axis=1)
+        expected = left.join(rdf)
+        tm.assert_frame_equal(result, expected)
+
+    def test_merge_na_keys(self):
+        data = [
+            [1950, "A", 1.5],
+            [1950, "B", 1.5],
+            [1955, "B", 1.5],
+            [1960, "B", np.nan],
+            [1970, "B", 4.0],
+            [1950, "C", 4.0],
+            [1960, "C", np.nan],
+            [1965, "C", 3.0],
+            [1970, "C", 4.0],
+        ]
+
+        frame = DataFrame(data, columns=["year", "panel", "data"])
+
+        other_data = [
+            [1960, "A", np.nan],
+            [1970, "A", np.nan],
+            [1955, "A", np.nan],
+            [1965, "A", np.nan],
+            [1965, "B", np.nan],
+            [1955, "C", np.nan],
+        ]
+        other = DataFrame(other_data, columns=["year", "panel", "data"])
+
+        result = frame.merge(other, how="outer")
+
+        expected = frame.fillna(-999).merge(other.fillna(-999), how="outer")
+        expected = expected.replace(-999, np.nan)
+
+        tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("klass", [None, np.asarray, Series, Index])
+    def test_merge_datetime_index(self, klass):
+        # see gh-19038
+        df = DataFrame(
+            [1, 2, 3], ["2016-01-01", "2017-01-01", "2018-01-01"], columns=["a"]
+        )
+        df.index = pd.to_datetime(df.index)
+        on_vector = df.index.year
+
+        if klass is not None:
+            on_vector = klass(on_vector)
+
+        expected = DataFrame(
+            OrderedDict([("a", [1, 2, 3]), ("key_1", [2016, 2017, 2018])])
+        )
+
+        result = df.merge(df, on=["a", on_vector], how="inner")
+        tm.assert_frame_equal(result, expected)
+
+        expected = DataFrame(
+            OrderedDict(
+                [("key_0", [2016, 2017, 2018]), ("a_x", [1, 2, 3]), ("a_y", [1, 2, 3])]
+            )
+        )
+
+        result = df.merge(df, on=[df.index.year], how="inner")
+        tm.assert_frame_equal(result, expected)
+
+    def test_join_multi_levels(self):
+
+        # GH 3662
+        # merge multi-levels
+        household = DataFrame(
+            dict(
+                household_id=[1, 2, 3],
+                male=[0, 1, 0],
+                wealth=[196087.3, 316478.7, 294750],
+            ),
+            columns=["household_id", "male", "wealth"],
+        ).set_index("household_id")
+        portfolio = DataFrame(
+            dict(
+                household_id=[1, 2, 2, 3, 3, 3, 4],
+                asset_id=[
+                    "nl0000301109",
+                    "nl0000289783",
+                    "gb00b03mlx29",
+                    "gb00b03mlx29",
+                    "lu0197800237",
+                    "nl0000289965",
+                    np.nan,
+                ],
+                name=[
+                    "ABN Amro",
+                    "Robeco",
+                    "Royal Dutch Shell",
+                    "Royal Dutch Shell",
+                    "AAB Eastern Europe Equity Fund",
+                    "Postbank BioTech Fonds",
+                    np.nan,
+                ],
+                share=[1.0, 0.4, 0.6, 0.15, 0.6, 0.25, 1.0],
+            ),
+            columns=["household_id", "asset_id", "name", "share"],
+        ).set_index(["household_id", "asset_id"])
+        result = household.join(portfolio, how="inner")
+        expected = (
+            DataFrame(
+                dict(
+                    male=[0, 1, 1, 0, 0, 0],
+                    wealth=[196087.3, 316478.7, 316478.7, 294750.0, 294750.0, 294750.0],
+                    name=[
+                        "ABN Amro",
+                        "Robeco",
+                        "Royal Dutch Shell",
+                        "Royal Dutch Shell",
+                        "AAB Eastern Europe Equity Fund",
+                        "Postbank BioTech Fonds",
+                    ],
+                    share=[1.00, 0.40, 0.60, 0.15, 0.60, 0.25],
+                    household_id=[1, 2, 2, 3, 3, 3],
+                    asset_id=[
+                        "nl0000301109",
+                        "nl0000289783",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "lu0197800237",
+                        "nl0000289965",
+                    ],
+                )
+            )
+            .set_index(["household_id", "asset_id"])
+            .reindex(columns=["male", "wealth", "name", "share"])
+        )
+        tm.assert_frame_equal(result, expected)
+
+        # equivalency
+        result = merge(
+            household.reset_index(),
+            portfolio.reset_index(),
+            on=["household_id"],
+            how="inner",
+        ).set_index(["household_id", "asset_id"])
+        tm.assert_frame_equal(result, expected)
+
+        result = household.join(portfolio, how="outer")
+        expected = concat(
+            [
+                expected,
+                (
+                    DataFrame(
+                        dict(share=[1.00]),
+                        index=MultiIndex.from_tuples(
+                            [(4, np.nan)], names=["household_id", "asset_id"]
+                        ),
+                    )
+                ),
+            ],
+            axis=0,
+            sort=True,
+        ).reindex(columns=expected.columns)
+        tm.assert_frame_equal(result, expected)
+
+        # invalid cases
+        household.index.name = "foo"
+
+        with pytest.raises(ValueError):
+            household.join(portfolio, how="inner")
+
+        portfolio2 = portfolio.copy()
+        portfolio2.index.set_names(["household_id", "foo"])
+
+        with pytest.raises(ValueError):
+            portfolio2.join(portfolio, how="inner")
+
+    def test_join_multi_levels2(self):
+
+        # some more advanced merges
+        # GH6360
+        household = DataFrame(
+            dict(
+                household_id=[1, 2, 2, 3, 3, 3, 4],
+                asset_id=[
+                    "nl0000301109",
+                    "nl0000301109",
+                    "gb00b03mlx29",
+                    "gb00b03mlx29",
+                    "lu0197800237",
+                    "nl0000289965",
+                    np.nan,
+                ],
+                share=[1.0, 0.4, 0.6, 0.15, 0.6, 0.25, 1.0],
+            ),
+            columns=["household_id", "asset_id", "share"],
+        ).set_index(["household_id", "asset_id"])
+
+        log_return = DataFrame(
+            dict(
+                asset_id=[
+                    "gb00b03mlx29",
+                    "gb00b03mlx29",
+                    "gb00b03mlx29",
+                    "lu0197800237",
+                    "lu0197800237",
+                ],
+                t=[233, 234, 235, 180, 181],
+                log_return=[0.09604978, -0.06524096, 0.03532373, 0.03025441, 0.036997],
+            )
+        ).set_index(["asset_id", "t"])
+
+        expected = (
+            DataFrame(
+                dict(
+                    household_id=[2, 2, 2, 3, 3, 3, 3, 3],
+                    asset_id=[
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "lu0197800237",
+                        "lu0197800237",
+                    ],
+                    t=[233, 234, 235, 233, 234, 235, 180, 181],
+                    share=[0.6, 0.6, 0.6, 0.15, 0.15, 0.15, 0.6, 0.6],
+                    log_return=[
+                        0.09604978,
+                        -0.06524096,
+                        0.03532373,
+                        0.09604978,
+                        -0.06524096,
+                        0.03532373,
+                        0.03025441,
+                        0.036997,
+                    ],
+                )
+            )
+            .set_index(["household_id", "asset_id", "t"])
+            .reindex(columns=["share", "log_return"])
+        )
+
+        # this is the equivalency
+        result = merge(
+            household.reset_index(),
+            log_return.reset_index(),
+            on=["asset_id"],
+            how="inner",
+        ).set_index(["household_id", "asset_id", "t"])
+        tm.assert_frame_equal(result, expected)
+
+        expected = (
+            DataFrame(
+                dict(
+                    household_id=[1, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 4],
+                    asset_id=[
+                        "nl0000301109",
+                        "nl0000301109",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "gb00b03mlx29",
+                        "lu0197800237",
+                        "lu0197800237",
+                        "nl0000289965",
+                        None,
+                    ],
+                    t=[None, None, 233, 234, 235, 233, 234, 235, 180, 181, None, None],
+                    share=[
+                        1.0,
+                        0.4,
+                        0.6,
+                        0.6,
+                        0.6,
+                        0.15,
+                        0.15,
+                        0.15,
+                        0.6,
+                        0.6,
+                        0.25,
+                        1.0,
+                    ],
+                    log_return=[
+                        None,
+                        None,
+                        0.09604978,
+                        -0.06524096,
+                        0.03532373,
+                        0.09604978,
+                        -0.06524096,
+                        0.03532373,
+                        0.03025441,
+                        0.036997,
+                        None,
+                        None,
+                    ],
+                )
+            )
+            .set_index(["household_id", "asset_id", "t"])
+            .reindex(columns=["share", "log_return"])
+        )
+
+        result = merge(
+            household.reset_index(),
+            log_return.reset_index(),
+            on=["asset_id"],
+            how="outer",
+        ).set_index(["household_id", "asset_id", "t"])
+
+        tm.assert_frame_equal(result, expected)
+
+
+class TestJoinMultiMulti:
+    def test_join_multi_multi(
+        self, left_multi, right_multi, join_type, on_cols_multi, idx_cols_multi
+    ):
+        # Multi-index join tests
+        expected = (
+            pd.merge(
+                left_multi.reset_index(),
+                right_multi.reset_index(),
+                how=join_type,
+                on=on_cols_multi,
+            )
+            .set_index(idx_cols_multi)
+            .sort_index()
+        )
+
+        result = left_multi.join(right_multi, how=join_type).sort_index()
+        tm.assert_frame_equal(result, expected)
+
+    def test_join_multi_empty_frames(
+        self, left_multi, right_multi, join_type, on_cols_multi, idx_cols_multi
+    ):
+
+        left_multi = left_multi.drop(columns=left_multi.columns)
+        right_multi = right_multi.drop(columns=right_multi.columns)
+
+        expected = (
+            pd.merge(
+                left_multi.reset_index(),
+                right_multi.reset_index(),
+                how=join_type,
+                on=on_cols_multi,
+            )
+            .set_index(idx_cols_multi)
+            .sort_index()
+        )
+
+        result = left_multi.join(right_multi, how=join_type).sort_index()
+        tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("box", [None, np.asarray, Series, Index])
+    def test_merge_datetime_index(self, box):
+        # see gh-19038
+        df = DataFrame(
+            [1, 2, 3], ["2016-01-01", "2017-01-01", "2018-01-01"], columns=["a"]
+        )
+        df.index = pd.to_datetime(df.index)
+        on_vector = df.index.year
+
+        if box is not None:
+            on_vector = box(on_vector)
+
+        expected = DataFrame(
+            OrderedDict([("a", [1, 2, 3]), ("key_1", [2016, 2017, 2018])])
+        )
+
+        result = df.merge(df, on=["a", on_vector], how="inner")
+        tm.assert_frame_equal(result, expected)
+
+        expected = DataFrame(
+            OrderedDict(
+                [("key_0", [2016, 2017, 2018]), ("a_x", [1, 2, 3]), ("a_y", [1, 2, 3])]
+            )
+        )
+
+        result = df.merge(df, on=[df.index.year], how="inner")
+        tm.assert_frame_equal(result, expected)
+
+    def test_single_common_level(self):
+        index_left = pd.MultiIndex.from_tuples(
+            [("K0", "X0"), ("K0", "X1"), ("K1", "X2")], names=["key", "X"]
+        )
+
+        left = pd.DataFrame(
+            {"A": ["A0", "A1", "A2"], "B": ["B0", "B1", "B2"]}, index=index_left
+        )
+
+        index_right = pd.MultiIndex.from_tuples(
+            [("K0", "Y0"), ("K1", "Y1"), ("K2", "Y2"), ("K2", "Y3")], names=["key", "Y"]
+        )
+
+        right = pd.DataFrame(
+            {"C": ["C0", "C1", "C2", "C3"], "D": ["D0", "D1", "D2", "D3"]},
+            index=index_right,
+        )
+
+        result = left.join(right)
+        expected = pd.merge(
+            left.reset_index(), right.reset_index(), on=["key"], how="inner"
+        ).set_index(["key", "X", "Y"])
+
+        tm.assert_frame_equal(result, expected)

venv/lib/python3.6/site-packages/pandas/tests/reshape/test_cut.py

@@ -0,0 +1,587 @@
+import numpy as np
+import pytest
+
+import pandas as pd
+from pandas import (
+    Categorical,
+    DataFrame,
+    DatetimeIndex,
+    Index,
+    Interval,
+    IntervalIndex,
+    Series,
+    TimedeltaIndex,
+    Timestamp,
+    cut,
+    date_range,
+    isna,
+    qcut,
+    timedelta_range,
+    to_datetime,
+)
+from pandas.api.types import CategoricalDtype as CDT
+import pandas.core.reshape.tile as tmod
+import pandas.util.testing as tm
+
+
+def test_simple():
+    data = np.ones(5, dtype="int64")
+    result = cut(data, 4, labels=False)
+
+    expected = np.array([1, 1, 1, 1, 1])
+    tm.assert_numpy_array_equal(result, expected, check_dtype=False)
+
+
+def test_bins():
+    data = np.array([0.2, 1.4, 2.5, 6.2, 9.7, 2.1])
+    result, bins = cut(data, 3, retbins=True)
+
+    intervals = IntervalIndex.from_breaks(bins.round(3))
+    intervals = intervals.take([0, 0, 0, 1, 2, 0])
+    expected = Categorical(intervals, ordered=True)
+
+    tm.assert_categorical_equal(result, expected)
+    tm.assert_almost_equal(bins, np.array([0.1905, 3.36666667, 6.53333333, 9.7]))
+
+
+def test_right():
+    data = np.array([0.2, 1.4, 2.5, 6.2, 9.7, 2.1, 2.575])
+    result, bins = cut(data, 4, right=True, retbins=True)
+
+    intervals = IntervalIndex.from_breaks(bins.round(3))
+    expected = Categorical(intervals, ordered=True)
+    expected = expected.take([0, 0, 0, 2, 3, 0, 0])
+
+    tm.assert_categorical_equal(result, expected)
+    tm.assert_almost_equal(bins, np.array([0.1905, 2.575, 4.95, 7.325, 9.7]))
+
+
+def test_no_right():
+    data = np.array([0.2, 1.4, 2.5, 6.2, 9.7, 2.1, 2.575])
+    result, bins = cut(data, 4, right=False, retbins=True)
+
+    intervals = IntervalIndex.from_breaks(bins.round(3), closed="left")
+    intervals = intervals.take([0, 0, 0, 2, 3, 0, 1])
+    expected = Categorical(intervals, ordered=True)
+
+    tm.assert_categorical_equal(result, expected)
+    tm.assert_almost_equal(bins, np.array([0.2, 2.575, 4.95, 7.325, 9.7095]))
+
+
+def test_array_like():
+    data = [0.2, 1.4, 2.5, 6.2, 9.7, 2.1]
+    result, bins = cut(data, 3, retbins=True)
+
+    intervals = IntervalIndex.from_breaks(bins.round(3))
+    intervals = intervals.take([0, 0, 0, 1, 2, 0])
+    expected = Categorical(intervals, ordered=True)
+
+    tm.assert_categorical_equal(result, expected)
+    tm.assert_almost_equal(bins, np.array([0.1905, 3.36666667, 6.53333333, 9.7]))
+
+
+def test_bins_from_interval_index():
+    c = cut(range(5), 3)
+    expected = c
+    result = cut(range(5), bins=expected.categories)
+    tm.assert_categorical_equal(result, expected)
+
+    expected = Categorical.from_codes(
+        np.append(c.codes, -1), categories=c.categories, ordered=True
+    )
+    result = cut(range(6), bins=expected.categories)
+    tm.assert_categorical_equal(result, expected)
+
+
+def test_bins_from_interval_index_doc_example():
+    # Make sure we preserve the bins.
+    ages = np.array([10, 15, 13, 12, 23, 25, 28, 59, 60])
+    c = cut(ages, bins=[0, 18, 35, 70])
+    expected = IntervalIndex.from_tuples([(0, 18), (18, 35), (35, 70)])
+    tm.assert_index_equal(c.categories, expected)
+
+    result = cut([25, 20, 50], bins=c.categories)
+    tm.assert_index_equal(result.categories, expected)
+    tm.assert_numpy_array_equal(result.codes, np.array([1, 1, 2], dtype="int8"))
+
+
+def test_bins_not_overlapping_from_interval_index():
+    # see gh-23980
+    msg = "Overlapping IntervalIndex is not accepted"
+    ii = IntervalIndex.from_tuples([(0, 10), (2, 12), (4, 14)])
+
+    with pytest.raises(ValueError, match=msg):
+        cut([5, 6], bins=ii)
+
+
+def test_bins_not_monotonic():
+    msg = "bins must increase monotonically"
+    data = [0.2, 1.4, 2.5, 6.2, 9.7, 2.1]
+
+    with pytest.raises(ValueError, match=msg):
+        cut(data, [0.1, 1.5, 1, 10])
+
+
+@pytest.mark.parametrize(
+    "x, bins, expected",
+    [
+        (
+            date_range("2017-12-31", periods=3),
+            [Timestamp.min, Timestamp("2018-01-01"), Timestamp.max],
+            IntervalIndex.from_tuples(
+                [
+                    (Timestamp.min, Timestamp("2018-01-01")),
+                    (Timestamp("2018-01-01"), Timestamp.max),
+                ]
+            ),
+        ),
+        (
+            [-1, 0, 1],
+            np.array(
+                [np.iinfo(np.int64).min, 0, np.iinfo(np.int64).max], dtype="int64"
+            ),
+            IntervalIndex.from_tuples(
+                [(np.iinfo(np.int64).min, 0), (0, np.iinfo(np.int64).max)]
+            ),
+        ),
+        (
+            [np.timedelta64(-1), np.timedelta64(0), np.timedelta64(1)],
+            np.array(
+                [
+                    np.timedelta64(-np.iinfo(np.int64).max),
+                    np.timedelta64(0),
+                    np.timedelta64(np.iinfo(np.int64).max),
+                ]
+            ),
+            IntervalIndex.from_tuples(
+                [
+                    (np.timedelta64(-np.iinfo(np.int64).max), np.timedelta64(0)),
+                    (np.timedelta64(0), np.timedelta64(np.iinfo(np.int64).max)),
+                ]
+            ),
+        ),
+    ],
+)
+def test_bins_monotonic_not_overflowing(x, bins, expected):
+    # GH 26045
+    result = cut(x, bins)
+    tm.assert_index_equal(result.categories, expected)
+
+
+def test_wrong_num_labels():
+    msg = "Bin labels must be one fewer than the number of bin edges"
+    data = [0.2, 1.4, 2.5, 6.2, 9.7, 2.1]
+
+    with pytest.raises(ValueError, match=msg):
+        cut(data, [0, 1, 10], labels=["foo", "bar", "baz"])
+
+
+@pytest.mark.parametrize(
+    "x,bins,msg",
+    [
+        ([], 2, "Cannot cut empty array"),
+        ([1, 2, 3], 0.5, "`bins` should be a positive integer"),
+    ],
+)
+def test_cut_corner(x, bins, msg):
+    with pytest.raises(ValueError, match=msg):
+        cut(x, bins)
+
+
+@pytest.mark.parametrize("arg", [2, np.eye(2), DataFrame(np.eye(2))])
+@pytest.mark.parametrize("cut_func", [cut, qcut])
+def test_cut_not_1d_arg(arg, cut_func):
+    msg = "Input array must be 1 dimensional"
+    with pytest.raises(ValueError, match=msg):
+        cut_func(arg, 2)
+
+
+@pytest.mark.parametrize(
+    "data",
+    [
+        [0, 1, 2, 3, 4, np.inf],
+        [-np.inf, 0, 1, 2, 3, 4],
+        [-np.inf, 0, 1, 2, 3, 4, np.inf],
+    ],
+)
+def test_int_bins_with_inf(data):
+    # GH 24314
+    msg = "cannot specify integer `bins` when input data contains infinity"
+    with pytest.raises(ValueError, match=msg):
+        cut(data, bins=3)
+
+
+def test_cut_out_of_range_more():
+    # see gh-1511
+    name = "x"
+
+    ser = Series([0, -1, 0, 1, -3], name=name)
+    ind = cut(ser, [0, 1], labels=False)
+
+    exp = Series([np.nan, np.nan, np.nan, 0, np.nan], name=name)
+    tm.assert_series_equal(ind, exp)
+
+
+@pytest.mark.parametrize(
+    "right,breaks,closed",
+    [
+        (True, [-1e-3, 0.25, 0.5, 0.75, 1], "right"),
+        (False, [0, 0.25, 0.5, 0.75, 1 + 1e-3], "left"),
+    ],
+)
+def test_labels(right, breaks, closed):
+    arr = np.tile(np.arange(0, 1.01, 0.1), 4)
+
+    result, bins = cut(arr, 4, retbins=True, right=right)
+    ex_levels = IntervalIndex.from_breaks(breaks, closed=closed)
+    tm.assert_index_equal(result.categories, ex_levels)
+
+
+def test_cut_pass_series_name_to_factor():
+    name = "foo"
+    ser = Series(np.random.randn(100), name=name)
+
+    factor = cut(ser, 4)
+    assert factor.name == name
+
+
+def test_label_precision():
+    arr = np.arange(0, 0.73, 0.01)
+    result = cut(arr, 4, precision=2)
+
+    ex_levels = IntervalIndex.from_breaks([-0.00072, 0.18, 0.36, 0.54, 0.72])
+    tm.assert_index_equal(result.categories, ex_levels)
+
+
+@pytest.mark.parametrize("labels", [None, False])
+def test_na_handling(labels):
+    arr = np.arange(0, 0.75, 0.01)
+    arr[::3] = np.nan
+
+    result = cut(arr, 4, labels=labels)
+    result = np.asarray(result)
+
+    expected = np.where(isna(arr), np.nan, result)
+    tm.assert_almost_equal(result, expected)
+
+
+def test_inf_handling():
+    data = np.arange(6)
+    data_ser = Series(data, dtype="int64")
+
+    bins = [-np.inf, 2, 4, np.inf]
+    result = cut(data, bins)
+    result_ser = cut(data_ser, bins)
+
+    ex_uniques = IntervalIndex.from_breaks(bins)
+    tm.assert_index_equal(result.categories, ex_uniques)
+
+    assert result[5] == Interval(4, np.inf)
+    assert result[0] == Interval(-np.inf, 2)
+    assert result_ser[5] == Interval(4, np.inf)
+    assert result_ser[0] == Interval(-np.inf, 2)
+
+
+def test_cut_out_of_bounds():
+    arr = np.random.randn(100)
+    result = cut(arr, [-1, 0, 1])
+
+    mask = isna(result)
+    ex_mask = (arr < -1) | (arr > 1)
+    tm.assert_numpy_array_equal(mask, ex_mask)
+
+
+@pytest.mark.parametrize(
+    "get_labels,get_expected",
+    [
+        (
+            lambda labels: labels,
+            lambda labels: Categorical(
+                ["Medium"] + 4 * ["Small"] + ["Medium", "Large"],
+                categories=labels,
+                ordered=True,
+            ),
+        ),
+        (
+            lambda labels: Categorical.from_codes([0, 1, 2], labels),
+            lambda labels: Categorical.from_codes([1] + 4 * [0] + [1, 2], labels),
+        ),
+    ],
+)
+def test_cut_pass_labels(get_labels, get_expected):
+    bins = [0, 25, 50, 100]
+    arr = [50, 5, 10, 15, 20, 30, 70]
+    labels = ["Small", "Medium", "Large"]
+
+    result = cut(arr, bins, labels=get_labels(labels))
+    tm.assert_categorical_equal(result, get_expected(labels))
+
+
+def test_cut_pass_labels_compat():
+    # see gh-16459
+    arr = [50, 5, 10, 15, 20, 30, 70]
+    labels = ["Good", "Medium", "Bad"]
+
+    result = cut(arr, 3, labels=labels)
+    exp = cut(arr, 3, labels=Categorical(labels, categories=labels, ordered=True))
+    tm.assert_categorical_equal(result, exp)
+
+
+@pytest.mark.parametrize("x", [np.arange(11.0), np.arange(11.0) / 1e10])
+def test_round_frac_just_works(x):
+    # It works.
+    cut(x, 2)
+
+
+@pytest.mark.parametrize(
+    "val,precision,expected",
+    [
+        (-117.9998, 3, -118),
+        (117.9998, 3, 118),
+        (117.9998, 2, 118),
+        (0.000123456, 2, 0.00012),
+    ],
+)
+def test_round_frac(val, precision, expected):
+    # see gh-1979
+    result = tmod._round_frac(val, precision=precision)
+    assert result == expected
+
+
+def test_cut_return_intervals():
+    ser = Series([0, 1, 2, 3, 4, 5, 6, 7, 8])
+    result = cut(ser, 3)
+
+    exp_bins = np.linspace(0, 8, num=4).round(3)
+    exp_bins[0] -= 0.008
+
+    expected = Series(
+        IntervalIndex.from_breaks(exp_bins, closed="right").take(
+            [0, 0, 0, 1, 1, 1, 2, 2, 2]
+        )
+    ).astype(CDT(ordered=True))
+    tm.assert_series_equal(result, expected)
+
+
+def test_series_ret_bins():
+    # see gh-8589
+    ser = Series(np.arange(4))
+    result, bins = cut(ser, 2, retbins=True)
+
+    expected = Series(
+        IntervalIndex.from_breaks([-0.003, 1.5, 3], closed="right").repeat(2)
+    ).astype(CDT(ordered=True))
+    tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "kwargs,msg",
+    [
+        (dict(duplicates="drop"), None),
+        (dict(), "Bin edges must be unique"),
+        (dict(duplicates="raise"), "Bin edges must be unique"),
+        (dict(duplicates="foo"), "invalid value for 'duplicates' parameter"),
+    ],
+)
+def test_cut_duplicates_bin(kwargs, msg):
+    # see gh-20947
+    bins = [0, 2, 4, 6, 10, 10]
+    values = Series(np.array([1, 3, 5, 7, 9]), index=["a", "b", "c", "d", "e"])
+
+    if msg is not None:
+        with pytest.raises(ValueError, match=msg):
+            cut(values, bins, **kwargs)
+    else:
+        result = cut(values, bins, **kwargs)
+        expected = cut(values, pd.unique(bins))
+        tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize("data", [9.0, -9.0, 0.0])
+@pytest.mark.parametrize("length", [1, 2])
+def test_single_bin(data, length):
+    # see gh-14652, gh-15428
+    ser = Series([data] * length)
+    result = cut(ser, 1, labels=False)
+
+    expected = Series([0] * length)
+    tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "array_1_writeable,array_2_writeable", [(True, True), (True, False), (False, False)]
+)
+def test_cut_read_only(array_1_writeable, array_2_writeable):
+    # issue 18773
+    array_1 = np.arange(0, 100, 10)
+    array_1.flags.writeable = array_1_writeable
+
+    array_2 = np.arange(0, 100, 10)
+    array_2.flags.writeable = array_2_writeable
+
+    hundred_elements = np.arange(100)
+    tm.assert_categorical_equal(
+        cut(hundred_elements, array_1), cut(hundred_elements, array_2)
+    )
+
+
+@pytest.mark.parametrize(
+    "conv",
+    [
+        lambda v: Timestamp(v),
+        lambda v: to_datetime(v),
+        lambda v: np.datetime64(v),
+        lambda v: Timestamp(v).to_pydatetime(),
+    ],
+)
+def test_datetime_bin(conv):
+    data = [np.datetime64("2012-12-13"), np.datetime64("2012-12-15")]
+    bin_data = ["2012-12-12", "2012-12-14", "2012-12-16"]
+
+    expected = Series(
+        IntervalIndex(
+            [
+                Interval(Timestamp(bin_data[0]), Timestamp(bin_data[1])),
+                Interval(Timestamp(bin_data[1]), Timestamp(bin_data[2])),
+            ]
+        )
+    ).astype(CDT(ordered=True))
+
+    bins = [conv(v) for v in bin_data]
+    result = Series(cut(data, bins=bins))
+    tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "data",
+    [
+        to_datetime(Series(["2013-01-01", "2013-01-02", "2013-01-03"])),
+        [
+            np.datetime64("2013-01-01"),
+            np.datetime64("2013-01-02"),
+            np.datetime64("2013-01-03"),
+        ],
+        np.array(
+            [
+                np.datetime64("2013-01-01"),
+                np.datetime64("2013-01-02"),
+                np.datetime64("2013-01-03"),
+            ]
+        ),
+        DatetimeIndex(["2013-01-01", "2013-01-02", "2013-01-03"]),
+    ],
+)
+def test_datetime_cut(data):
+    # see gh-14714
+    #
+    # Testing time data when it comes in various collection types.
+    result, _ = cut(data, 3, retbins=True)
+    expected = Series(
+        IntervalIndex(
+            [
+                Interval(
+                    Timestamp("2012-12-31 23:57:07.200000"),
+                    Timestamp("2013-01-01 16:00:00"),
+                ),
+                Interval(
+                    Timestamp("2013-01-01 16:00:00"), Timestamp("2013-01-02 08:00:00")
+                ),
+                Interval(
+                    Timestamp("2013-01-02 08:00:00"), Timestamp("2013-01-03 00:00:00")
+                ),
+            ]
+        )
+    ).astype(CDT(ordered=True))
+    tm.assert_series_equal(Series(result), expected)
+
+
+@pytest.mark.parametrize(
+    "bins",
+    [
+        3,
+        [
+            Timestamp("2013-01-01 04:57:07.200000"),
+            Timestamp("2013-01-01 21:00:00"),
+            Timestamp("2013-01-02 13:00:00"),
+            Timestamp("2013-01-03 05:00:00"),
+        ],
+    ],
+)
+@pytest.mark.parametrize("box", [list, np.array, Index, Series])
+def test_datetime_tz_cut(bins, box):
+    # see gh-19872
+    tz = "US/Eastern"
+    s = Series(date_range("20130101", periods=3, tz=tz))
+
+    if not isinstance(bins, int):
+        bins = box(bins)
+
+    result = cut(s, bins)
+    expected = Series(
+        IntervalIndex(
+            [
+                Interval(
+                    Timestamp("2012-12-31 23:57:07.200000", tz=tz),
+                    Timestamp("2013-01-01 16:00:00", tz=tz),
+                ),
+                Interval(
+                    Timestamp("2013-01-01 16:00:00", tz=tz),
+                    Timestamp("2013-01-02 08:00:00", tz=tz),
+                ),
+                Interval(
+                    Timestamp("2013-01-02 08:00:00", tz=tz),
+                    Timestamp("2013-01-03 00:00:00", tz=tz),
+                ),
+            ]
+        )
+    ).astype(CDT(ordered=True))
+    tm.assert_series_equal(result, expected)
+
+
+def test_datetime_nan_error():
+    msg = "bins must be of datetime64 dtype"
+
+    with pytest.raises(ValueError, match=msg):
+        cut(date_range("20130101", periods=3), bins=[0, 2, 4])
+
+
+def test_datetime_nan_mask():
+    result = cut(
+        date_range("20130102", periods=5), bins=date_range("20130101", periods=2)
+    )
+
+    mask = result.categories.isna()
+    tm.assert_numpy_array_equal(mask, np.array([False]))
+
+    mask = result.isna()
+    tm.assert_numpy_array_equal(mask, np.array([False, True, True, True, True]))
+
+
+@pytest.mark.parametrize("tz", [None, "UTC", "US/Pacific"])
+def test_datetime_cut_roundtrip(tz):
+    # see gh-19891
+    ser = Series(date_range("20180101", periods=3, tz=tz))
+    result, result_bins = cut(ser, 2, retbins=True)
+
+    expected = cut(ser, result_bins)
+    tm.assert_series_equal(result, expected)
+
+    expected_bins = DatetimeIndex(
+        ["2017-12-31 23:57:07.200000", "2018-01-02 00:00:00", "2018-01-03 00:00:00"]
+    )
+    expected_bins = expected_bins.tz_localize(tz)
+    tm.assert_index_equal(result_bins, expected_bins)
+
+
+def test_timedelta_cut_roundtrip():
+    # see gh-19891
+    ser = Series(timedelta_range("1day", periods=3))
+    result, result_bins = cut(ser, 2, retbins=True)
+
+    expected = cut(ser, result_bins)
+    tm.assert_series_equal(result, expected)
+
+    expected_bins = TimedeltaIndex(
+        ["0 days 23:57:07.200000", "2 days 00:00:00", "3 days 00:00:00"]
+    )
+    tm.assert_index_equal(result_bins, expected_bins)

venv/lib/python3.6/site-packages/pandas/tests/reshape/test_melt.py

@@ -0,0 +1,964 @@
+import numpy as np
+from numpy import nan
+import pytest
+
+import pandas as pd
+from pandas import DataFrame, lreshape, melt, wide_to_long
+import pandas.util.testing as tm
+
+
+class TestMelt:
+    def setup_method(self, method):
+        self.df = tm.makeTimeDataFrame()[:10]
+        self.df["id1"] = (self.df["A"] > 0).astype(np.int64)
+        self.df["id2"] = (self.df["B"] > 0).astype(np.int64)
+
+        self.var_name = "var"
+        self.value_name = "val"
+
+        self.df1 = pd.DataFrame(
+            [
+                [1.067683, -1.110463, 0.20867],
+                [-1.321405, 0.368915, -1.055342],
+                [-0.807333, 0.08298, -0.873361],
+            ]
+        )
+        self.df1.columns = [list("ABC"), list("abc")]
+        self.df1.columns.names = ["CAP", "low"]
+
+    def test_top_level_method(self):
+        result = melt(self.df)
+        assert result.columns.tolist() == ["variable", "value"]
+
+    def test_method_signatures(self):
+        tm.assert_frame_equal(self.df.melt(), melt(self.df))
+
+        tm.assert_frame_equal(
+            self.df.melt(id_vars=["id1", "id2"], value_vars=["A", "B"]),
+            melt(self.df, id_vars=["id1", "id2"], value_vars=["A", "B"]),
+        )
+
+        tm.assert_frame_equal(
+            self.df.melt(var_name=self.var_name, value_name=self.value_name),
+            melt(self.df, var_name=self.var_name, value_name=self.value_name),
+        )
+
+        tm.assert_frame_equal(self.df1.melt(col_level=0), melt(self.df1, col_level=0))
+
+    def test_default_col_names(self):
+        result = self.df.melt()
+        assert result.columns.tolist() == ["variable", "value"]
+
+        result1 = self.df.melt(id_vars=["id1"])
+        assert result1.columns.tolist() == ["id1", "variable", "value"]
+
+        result2 = self.df.melt(id_vars=["id1", "id2"])
+        assert result2.columns.tolist() == ["id1", "id2", "variable", "value"]
+
+    def test_value_vars(self):
+        result3 = self.df.melt(id_vars=["id1", "id2"], value_vars="A")
+        assert len(result3) == 10
+
+        result4 = self.df.melt(id_vars=["id1", "id2"], value_vars=["A", "B"])
+        expected4 = DataFrame(
+            {
+                "id1": self.df["id1"].tolist() * 2,
+                "id2": self.df["id2"].tolist() * 2,
+                "variable": ["A"] * 10 + ["B"] * 10,
+                "value": (self.df["A"].tolist() + self.df["B"].tolist()),
+            },
+            columns=["id1", "id2", "variable", "value"],
+        )
+        tm.assert_frame_equal(result4, expected4)
+
+    def test_value_vars_types(self):
+        # GH 15348
+        expected = DataFrame(
+            {
+                "id1": self.df["id1"].tolist() * 2,
+                "id2": self.df["id2"].tolist() * 2,
+                "variable": ["A"] * 10 + ["B"] * 10,
+                "value": (self.df["A"].tolist() + self.df["B"].tolist()),
+            },
+            columns=["id1", "id2", "variable", "value"],
+        )
+
+        for type_ in (tuple, list, np.array):
+            result = self.df.melt(id_vars=["id1", "id2"], value_vars=type_(("A", "B")))
+            tm.assert_frame_equal(result, expected)
+
+    def test_vars_work_with_multiindex(self):
+        expected = DataFrame(
+            {
+                ("A", "a"): self.df1[("A", "a")],
+                "CAP": ["B"] * len(self.df1),
+                "low": ["b"] * len(self.df1),
+                "value": self.df1[("B", "b")],
+            },
+            columns=[("A", "a"), "CAP", "low", "value"],
+        )
+
+        result = self.df1.melt(id_vars=[("A", "a")], value_vars=[("B", "b")])
+        tm.assert_frame_equal(result, expected)
+
+    def test_single_vars_work_with_multiindex(self):
+        expected = DataFrame(
+            {
+                "A": {0: 1.067683, 1: -1.321405, 2: -0.807333},
+                "CAP": {0: "B", 1: "B", 2: "B"},
+                "value": {0: -1.110463, 1: 0.368915, 2: 0.08298},
+            }
+        )
+        result = self.df1.melt(["A"], ["B"], col_level=0)
+        tm.assert_frame_equal(result, expected)
+
+    def test_tuple_vars_fail_with_multiindex(self):
+        # melt should fail with an informative error message if
+        # the columns have a MultiIndex and a tuple is passed
+        # for id_vars or value_vars.
+        tuple_a = ("A", "a")
+        list_a = [tuple_a]
+        tuple_b = ("B", "b")
+        list_b = [tuple_b]
+
+        msg = r"(id|value)_vars must be a list of tuples when columns are a MultiIndex"
+        for id_vars, value_vars in (
+            (tuple_a, list_b),
+            (list_a, tuple_b),
+            (tuple_a, tuple_b),
+        ):
+            with pytest.raises(ValueError, match=msg):
+                self.df1.melt(id_vars=id_vars, value_vars=value_vars)
+
+    def test_custom_var_name(self):
+        result5 = self.df.melt(var_name=self.var_name)
+        assert result5.columns.tolist() == ["var", "value"]
+
+        result6 = self.df.melt(id_vars=["id1"], var_name=self.var_name)
+        assert result6.columns.tolist() == ["id1", "var", "value"]
+
+        result7 = self.df.melt(id_vars=["id1", "id2"], var_name=self.var_name)
+        assert result7.columns.tolist() == ["id1", "id2", "var", "value"]
+
+        result8 = self.df.melt(
+            id_vars=["id1", "id2"], value_vars="A", var_name=self.var_name
+        )
+        assert result8.columns.tolist() == ["id1", "id2", "var", "value"]
+
+        result9 = self.df.melt(
+            id_vars=["id1", "id2"], value_vars=["A", "B"], var_name=self.var_name
+        )
+        expected9 = DataFrame(
+            {
+                "id1": self.df["id1"].tolist() * 2,
+                "id2": self.df["id2"].tolist() * 2,
+                self.var_name: ["A"] * 10 + ["B"] * 10,
+                "value": (self.df["A"].tolist() + self.df["B"].tolist()),
+            },
+            columns=["id1", "id2", self.var_name, "value"],
+        )
+        tm.assert_frame_equal(result9, expected9)
+
+    def test_custom_value_name(self):
+        result10 = self.df.melt(value_name=self.value_name)
+        assert result10.columns.tolist() == ["variable", "val"]
+
+        result11 = self.df.melt(id_vars=["id1"], value_name=self.value_name)
+        assert result11.columns.tolist() == ["id1", "variable", "val"]
+
+        result12 = self.df.melt(id_vars=["id1", "id2"], value_name=self.value_name)
+        assert result12.columns.tolist() == ["id1", "id2", "variable", "val"]
+
+        result13 = self.df.melt(
+            id_vars=["id1", "id2"], value_vars="A", value_name=self.value_name
+        )
+        assert result13.columns.tolist() == ["id1", "id2", "variable", "val"]
+
+        result14 = self.df.melt(
+            id_vars=["id1", "id2"], value_vars=["A", "B"], value_name=self.value_name
+        )
+        expected14 = DataFrame(
+            {
+                "id1": self.df["id1"].tolist() * 2,
+                "id2": self.df["id2"].tolist() * 2,
+                "variable": ["A"] * 10 + ["B"] * 10,
+                self.value_name: (self.df["A"].tolist() + self.df["B"].tolist()),
+            },
+            columns=["id1", "id2", "variable", self.value_name],
+        )
+        tm.assert_frame_equal(result14, expected14)
+
+    def test_custom_var_and_value_name(self):
+
+        result15 = self.df.melt(var_name=self.var_name, value_name=self.value_name)
+        assert result15.columns.tolist() == ["var", "val"]
+
+        result16 = self.df.melt(
+            id_vars=["id1"], var_name=self.var_name, value_name=self.value_name
+        )
+        assert result16.columns.tolist() == ["id1", "var", "val"]
+
+        result17 = self.df.melt(
+            id_vars=["id1", "id2"], var_name=self.var_name, value_name=self.value_name
+        )
+        assert result17.columns.tolist() == ["id1", "id2", "var", "val"]
+
+        result18 = self.df.melt(
+            id_vars=["id1", "id2"],
+            value_vars="A",
+            var_name=self.var_name,
+            value_name=self.value_name,
+        )
+        assert result18.columns.tolist() == ["id1", "id2", "var", "val"]
+
+        result19 = self.df.melt(
+            id_vars=["id1", "id2"],
+            value_vars=["A", "B"],
+            var_name=self.var_name,
+            value_name=self.value_name,
+        )
+        expected19 = DataFrame(
+            {
+                "id1": self.df["id1"].tolist() * 2,
+                "id2": self.df["id2"].tolist() * 2,
+                self.var_name: ["A"] * 10 + ["B"] * 10,
+                self.value_name: (self.df["A"].tolist() + self.df["B"].tolist()),
+            },
+            columns=["id1", "id2", self.var_name, self.value_name],
+        )
+        tm.assert_frame_equal(result19, expected19)
+
+        df20 = self.df.copy()
+        df20.columns.name = "foo"
+        result20 = df20.melt()
+        assert result20.columns.tolist() == ["foo", "value"]
+
+    def test_col_level(self):
+        res1 = self.df1.melt(col_level=0)
+        res2 = self.df1.melt(col_level="CAP")
+        assert res1.columns.tolist() == ["CAP", "value"]
+        assert res2.columns.tolist() == ["CAP", "value"]
+
+    def test_multiindex(self):
+        res = self.df1.melt()
+        assert res.columns.tolist() == ["CAP", "low", "value"]
+
+    @pytest.mark.parametrize(
+        "col",
+        [
+            pd.Series(pd.date_range("2010", periods=5, tz="US/Pacific")),
+            pd.Series(["a", "b", "c", "a", "d"], dtype="category"),
+            pd.Series([0, 1, 0, 0, 0]),
+        ],
+    )
+    def test_pandas_dtypes(self, col):
+        # GH 15785
+        df = DataFrame(
+            {"klass": range(5), "col": col, "attr1": [1, 0, 0, 0, 0], "attr2": col}
+        )
+        expected_value = pd.concat([pd.Series([1, 0, 0, 0, 0]), col], ignore_index=True)
+        result = melt(
+            df, id_vars=["klass", "col"], var_name="attribute", value_name="value"
+        )
+        expected = DataFrame(
+            {
+                0: list(range(5)) * 2,
+                1: pd.concat([col] * 2, ignore_index=True),
+                2: ["attr1"] * 5 + ["attr2"] * 5,
+                3: expected_value,
+            }
+        )
+        expected.columns = ["klass", "col", "attribute", "value"]
+        tm.assert_frame_equal(result, expected)
+
+    def test_melt_missing_columns_raises(self):
+        # GH-23575
+        # This test is to ensure that pandas raises an error if melting is
+        # attempted with column names absent from the dataframe
+
+        # Generate data
+        df = pd.DataFrame(np.random.randn(5, 4), columns=list("abcd"))
+
+        # Try to melt with missing `value_vars` column name
+        msg = "The following '{Var}' are not present in the DataFrame: {Col}"
+        with pytest.raises(
+            KeyError, match=msg.format(Var="value_vars", Col="\\['C'\\]")
+        ):
+            df.melt(["a", "b"], ["C", "d"])
+
+        # Try to melt with missing `id_vars` column name
+        with pytest.raises(KeyError, match=msg.format(Var="id_vars", Col="\\['A'\\]")):
+            df.melt(["A", "b"], ["c", "d"])
+
+        # Multiple missing
+        with pytest.raises(
+            KeyError,
+            match=msg.format(Var="id_vars", Col="\\['not_here', 'or_there'\\]"),
+        ):
+            df.melt(["a", "b", "not_here", "or_there"], ["c", "d"])
+
+        # Multiindex melt fails if column is missing from multilevel melt
+        multi = df.copy()
+        multi.columns = [list("ABCD"), list("abcd")]
+        with pytest.raises(KeyError, match=msg.format(Var="id_vars", Col="\\['E'\\]")):
+            multi.melt([("E", "a")], [("B", "b")])
+        # Multiindex fails if column is missing from single level melt
+        with pytest.raises(
+            KeyError, match=msg.format(Var="value_vars", Col="\\['F'\\]")
+        ):
+            multi.melt(["A"], ["F"], col_level=0)
+
+
+class TestLreshape:
+    def test_pairs(self):
+        data = {
+            "birthdt": [
+                "08jan2009",
+                "20dec2008",
+                "30dec2008",
+                "21dec2008",
+                "11jan2009",
+            ],
+            "birthwt": [1766, 3301, 1454, 3139, 4133],
+            "id": [101, 102, 103, 104, 105],
+            "sex": ["Male", "Female", "Female", "Female", "Female"],
+            "visitdt1": [
+                "11jan2009",
+                "22dec2008",
+                "04jan2009",
+                "29dec2008",
+                "20jan2009",
+            ],
+            "visitdt2": ["21jan2009", nan, "22jan2009", "31dec2008", "03feb2009"],
+            "visitdt3": ["05feb2009", nan, nan, "02jan2009", "15feb2009"],
+            "wt1": [1823, 3338, 1549, 3298, 4306],
+            "wt2": [2011.0, nan, 1892.0, 3338.0, 4575.0],
+            "wt3": [2293.0, nan, nan, 3377.0, 4805.0],
+        }
+
+        df = DataFrame(data)
+
+        spec = {
+            "visitdt": ["visitdt{i:d}".format(i=i) for i in range(1, 4)],
+            "wt": ["wt{i:d}".format(i=i) for i in range(1, 4)],
+        }
+        result = lreshape(df, spec)
+
+        exp_data = {
+            "birthdt": [
+                "08jan2009",
+                "20dec2008",
+                "30dec2008",
+                "21dec2008",
+                "11jan2009",
+                "08jan2009",
+                "30dec2008",
+                "21dec2008",
+                "11jan2009",
+                "08jan2009",
+                "21dec2008",
+                "11jan2009",
+            ],
+            "birthwt": [
+                1766,
+                3301,
+                1454,
+                3139,
+                4133,
+                1766,
+                1454,
+                3139,
+                4133,
+                1766,
+                3139,
+                4133,
+            ],
+            "id": [101, 102, 103, 104, 105, 101, 103, 104, 105, 101, 104, 105],
+            "sex": [
+                "Male",
+                "Female",
+                "Female",
+                "Female",
+                "Female",
+                "Male",
+                "Female",
+                "Female",
+                "Female",
+                "Male",
+                "Female",
+                "Female",
+            ],
+            "visitdt": [
+                "11jan2009",
+                "22dec2008",
+                "04jan2009",
+                "29dec2008",
+                "20jan2009",
+                "21jan2009",
+                "22jan2009",
+                "31dec2008",
+                "03feb2009",
+                "05feb2009",
+                "02jan2009",
+                "15feb2009",
+            ],
+            "wt": [
+                1823.0,
+                3338.0,
+                1549.0,
+                3298.0,
+                4306.0,
+                2011.0,
+                1892.0,
+                3338.0,
+                4575.0,
+                2293.0,
+                3377.0,
+                4805.0,
+            ],
+        }
+        exp = DataFrame(exp_data, columns=result.columns)
+        tm.assert_frame_equal(result, exp)
+
+        result = lreshape(df, spec, dropna=False)
+        exp_data = {
+            "birthdt": [
+                "08jan2009",
+                "20dec2008",
+                "30dec2008",
+                "21dec2008",
+                "11jan2009",
+                "08jan2009",
+                "20dec2008",
+                "30dec2008",
+                "21dec2008",
+                "11jan2009",
+                "08jan2009",
+                "20dec2008",
+                "30dec2008",
+                "21dec2008",
+                "11jan2009",
+            ],
+            "birthwt": [
+                1766,
+                3301,
+                1454,
+                3139,
+                4133,
+                1766,
+                3301,
+                1454,
+                3139,
+                4133,
+                1766,
+                3301,
+                1454,
+                3139,
+                4133,
+            ],
+            "id": [
+                101,
+                102,
+                103,
+                104,
+                105,
+                101,
+                102,
+                103,
+                104,
+                105,
+                101,
+                102,
+                103,
+                104,
+                105,
+            ],
+            "sex": [
+                "Male",
+                "Female",
+                "Female",
+                "Female",
+                "Female",
+                "Male",
+                "Female",
+                "Female",
+                "Female",
+                "Female",
+                "Male",
+                "Female",
+                "Female",
+                "Female",
+                "Female",
+            ],
+            "visitdt": [
+                "11jan2009",
+                "22dec2008",
+                "04jan2009",
+                "29dec2008",
+                "20jan2009",
+                "21jan2009",
+                nan,
+                "22jan2009",
+                "31dec2008",
+                "03feb2009",
+                "05feb2009",
+                nan,
+                nan,
+                "02jan2009",
+                "15feb2009",
+            ],
+            "wt": [
+                1823.0,
+                3338.0,
+                1549.0,
+                3298.0,
+                4306.0,
+                2011.0,
+                nan,
+                1892.0,
+                3338.0,
+                4575.0,
+                2293.0,
+                nan,
+                nan,
+                3377.0,
+                4805.0,
+            ],
+        }
+        exp = DataFrame(exp_data, columns=result.columns)
+        tm.assert_frame_equal(result, exp)
+
+        spec = {
+            "visitdt": ["visitdt{i:d}".format(i=i) for i in range(1, 3)],
+            "wt": ["wt{i:d}".format(i=i) for i in range(1, 4)],
+        }
+        msg = "All column lists must be same length"
+        with pytest.raises(ValueError, match=msg):
+            lreshape(df, spec)
+
+
+class TestWideToLong:
+    def test_simple(self):
+        np.random.seed(123)
+        x = np.random.randn(3)
+        df = pd.DataFrame(
+            {
+                "A1970": {0: "a", 1: "b", 2: "c"},
+                "A1980": {0: "d", 1: "e", 2: "f"},
+                "B1970": {0: 2.5, 1: 1.2, 2: 0.7},
+                "B1980": {0: 3.2, 1: 1.3, 2: 0.1},
+                "X": dict(zip(range(3), x)),
+            }
+        )
+        df["id"] = df.index
+        exp_data = {
+            "X": x.tolist() + x.tolist(),
+            "A": ["a", "b", "c", "d", "e", "f"],
+            "B": [2.5, 1.2, 0.7, 3.2, 1.3, 0.1],
+            "year": [1970, 1970, 1970, 1980, 1980, 1980],
+            "id": [0, 1, 2, 0, 1, 2],
+        }
+        expected = DataFrame(exp_data)
+        expected = expected.set_index(["id", "year"])[["X", "A", "B"]]
+        result = wide_to_long(df, ["A", "B"], i="id", j="year")
+        tm.assert_frame_equal(result, expected)
+
+    def test_stubs(self):
+        # GH9204
+        df = pd.DataFrame([[0, 1, 2, 3, 8], [4, 5, 6, 7, 9]])
+        df.columns = ["id", "inc1", "inc2", "edu1", "edu2"]
+        stubs = ["inc", "edu"]
+
+        # TODO: unused?
+        df_long = pd.wide_to_long(df, stubs, i="id", j="age")  # noqa
+
+        assert stubs == ["inc", "edu"]
+
+    def test_separating_character(self):
+        # GH14779
+        np.random.seed(123)
+        x = np.random.randn(3)
+        df = pd.DataFrame(
+            {
+                "A.1970": {0: "a", 1: "b", 2: "c"},
+                "A.1980": {0: "d", 1: "e", 2: "f"},
+                "B.1970": {0: 2.5, 1: 1.2, 2: 0.7},
+                "B.1980": {0: 3.2, 1: 1.3, 2: 0.1},
+                "X": dict(zip(range(3), x)),
+            }
+        )
+        df["id"] = df.index
+        exp_data = {
+            "X": x.tolist() + x.tolist(),
+            "A": ["a", "b", "c", "d", "e", "f"],
+            "B": [2.5, 1.2, 0.7, 3.2, 1.3, 0.1],
+            "year": [1970, 1970, 1970, 1980, 1980, 1980],
+            "id": [0, 1, 2, 0, 1, 2],
+        }
+        expected = DataFrame(exp_data)
+        expected = expected.set_index(["id", "year"])[["X", "A", "B"]]
+        result = wide_to_long(df, ["A", "B"], i="id", j="year", sep=".")
+        tm.assert_frame_equal(result, expected)
+
+    def test_escapable_characters(self):
+        np.random.seed(123)
+        x = np.random.randn(3)
+        df = pd.DataFrame(
+            {
+                "A(quarterly)1970": {0: "a", 1: "b", 2: "c"},
+                "A(quarterly)1980": {0: "d", 1: "e", 2: "f"},
+                "B(quarterly)1970": {0: 2.5, 1: 1.2, 2: 0.7},
+                "B(quarterly)1980": {0: 3.2, 1: 1.3, 2: 0.1},
+                "X": dict(zip(range(3), x)),
+            }
+        )
+        df["id"] = df.index
+        exp_data = {
+            "X": x.tolist() + x.tolist(),
+            "A(quarterly)": ["a", "b", "c", "d", "e", "f"],
+            "B(quarterly)": [2.5, 1.2, 0.7, 3.2, 1.3, 0.1],
+            "year": [1970, 1970, 1970, 1980, 1980, 1980],
+            "id": [0, 1, 2, 0, 1, 2],
+        }
+        expected = DataFrame(exp_data)
+        expected = expected.set_index(["id", "year"])[
+            ["X", "A(quarterly)", "B(quarterly)"]
+        ]
+        result = wide_to_long(df, ["A(quarterly)", "B(quarterly)"], i="id", j="year")
+        tm.assert_frame_equal(result, expected)
+
+    def test_unbalanced(self):
+        # test that we can have a varying amount of time variables
+        df = pd.DataFrame(
+            {
+                "A2010": [1.0, 2.0],
+                "A2011": [3.0, 4.0],
+                "B2010": [5.0, 6.0],
+                "X": ["X1", "X2"],
+            }
+        )
+        df["id"] = df.index
+        exp_data = {
+            "X": ["X1", "X1", "X2", "X2"],
+            "A": [1.0, 3.0, 2.0, 4.0],
+            "B": [5.0, np.nan, 6.0, np.nan],
+            "id": [0, 0, 1, 1],
+            "year": [2010, 2011, 2010, 2011],
+        }
+        expected = pd.DataFrame(exp_data)
+        expected = expected.set_index(["id", "year"])[["X", "A", "B"]]
+        result = wide_to_long(df, ["A", "B"], i="id", j="year")
+        tm.assert_frame_equal(result, expected)
+
+    def test_character_overlap(self):
+        # Test we handle overlapping characters in both id_vars and value_vars
+        df = pd.DataFrame(
+            {
+                "A11": ["a11", "a22", "a33"],
+                "A12": ["a21", "a22", "a23"],
+                "B11": ["b11", "b12", "b13"],
+                "B12": ["b21", "b22", "b23"],
+                "BB11": [1, 2, 3],
+                "BB12": [4, 5, 6],
+                "BBBX": [91, 92, 93],
+                "BBBZ": [91, 92, 93],
+            }
+        )
+        df["id"] = df.index
+        expected = pd.DataFrame(
+            {
+                "BBBX": [91, 92, 93, 91, 92, 93],
+                "BBBZ": [91, 92, 93, 91, 92, 93],
+                "A": ["a11", "a22", "a33", "a21", "a22", "a23"],
+                "B": ["b11", "b12", "b13", "b21", "b22", "b23"],
+                "BB": [1, 2, 3, 4, 5, 6],
+                "id": [0, 1, 2, 0, 1, 2],
+                "year": [11, 11, 11, 12, 12, 12],
+            }
+        )
+        expected = expected.set_index(["id", "year"])[["BBBX", "BBBZ", "A", "B", "BB"]]
+        result = wide_to_long(df, ["A", "B", "BB"], i="id", j="year")
+        tm.assert_frame_equal(result.sort_index(axis=1), expected.sort_index(axis=1))
+
+    def test_invalid_separator(self):
+        # if an invalid separator is supplied a empty data frame is returned
+        sep = "nope!"
+        df = pd.DataFrame(
+            {
+                "A2010": [1.0, 2.0],
+                "A2011": [3.0, 4.0],
+                "B2010": [5.0, 6.0],
+                "X": ["X1", "X2"],
+            }
+        )
+        df["id"] = df.index
+        exp_data = {
+            "X": "",
+            "A2010": [],
+            "A2011": [],
+            "B2010": [],
+            "id": [],
+            "year": [],
+            "A": [],
+            "B": [],
+        }
+        expected = pd.DataFrame(exp_data).astype({"year": "int"})
+        expected = expected.set_index(["id", "year"])[
+            ["X", "A2010", "A2011", "B2010", "A", "B"]
+        ]
+        expected.index.set_levels([0, 1], level=0, inplace=True)
+        result = wide_to_long(df, ["A", "B"], i="id", j="year", sep=sep)
+        tm.assert_frame_equal(result.sort_index(axis=1), expected.sort_index(axis=1))
+
+    def test_num_string_disambiguation(self):
+        # Test that we can disambiguate number value_vars from
+        # string value_vars
+        df = pd.DataFrame(
+            {
+                "A11": ["a11", "a22", "a33"],
+                "A12": ["a21", "a22", "a23"],
+                "B11": ["b11", "b12", "b13"],
+                "B12": ["b21", "b22", "b23"],
+                "BB11": [1, 2, 3],
+                "BB12": [4, 5, 6],
+                "Arating": [91, 92, 93],
+                "Arating_old": [91, 92, 93],
+            }
+        )
+        df["id"] = df.index
+        expected = pd.DataFrame(
+            {
+                "Arating": [91, 92, 93, 91, 92, 93],
+                "Arating_old": [91, 92, 93, 91, 92, 93],
+                "A": ["a11", "a22", "a33", "a21", "a22", "a23"],
+                "B": ["b11", "b12", "b13", "b21", "b22", "b23"],
+                "BB": [1, 2, 3, 4, 5, 6],
+                "id": [0, 1, 2, 0, 1, 2],
+                "year": [11, 11, 11, 12, 12, 12],
+            }
+        )
+        expected = expected.set_index(["id", "year"])[
+            ["Arating", "Arating_old", "A", "B", "BB"]
+        ]
+        result = wide_to_long(df, ["A", "B", "BB"], i="id", j="year")
+        tm.assert_frame_equal(result.sort_index(axis=1), expected.sort_index(axis=1))
+
+    def test_invalid_suffixtype(self):
+        # If all stubs names end with a string, but a numeric suffix is
+        # assumed,  an empty data frame is returned
+        df = pd.DataFrame(
+            {
+                "Aone": [1.0, 2.0],
+                "Atwo": [3.0, 4.0],
+                "Bone": [5.0, 6.0],
+                "X": ["X1", "X2"],
+            }
+        )
+        df["id"] = df.index
+        exp_data = {
+            "X": "",
+            "Aone": [],
+            "Atwo": [],
+            "Bone": [],
+            "id": [],
+            "year": [],
+            "A": [],
+            "B": [],
+        }
+        expected = pd.DataFrame(exp_data).astype({"year": "int"})
+
+        expected = expected.set_index(["id", "year"])
+        expected.index.set_levels([0, 1], level=0, inplace=True)
+        result = wide_to_long(df, ["A", "B"], i="id", j="year")
+        tm.assert_frame_equal(result.sort_index(axis=1), expected.sort_index(axis=1))
+
+    def test_multiple_id_columns(self):
+        # Taken from http://www.ats.ucla.edu/stat/stata/modules/reshapel.htm
+        df = pd.DataFrame(
+            {
+                "famid": [1, 1, 1, 2, 2, 2, 3, 3, 3],
+                "birth": [1, 2, 3, 1, 2, 3, 1, 2, 3],
+                "ht1": [2.8, 2.9, 2.2, 2, 1.8, 1.9, 2.2, 2.3, 2.1],
+                "ht2": [3.4, 3.8, 2.9, 3.2, 2.8, 2.4, 3.3, 3.4, 2.9],
+            }
+        )
+        expected = pd.DataFrame(
+            {
+                "ht": [
+                    2.8,
+                    3.4,
+                    2.9,
+                    3.8,
+                    2.2,
+                    2.9,
+                    2.0,
+                    3.2,
+                    1.8,
+                    2.8,
+                    1.9,
+                    2.4,
+                    2.2,
+                    3.3,
+                    2.3,
+                    3.4,
+                    2.1,
+                    2.9,
+                ],
+                "famid": [1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3],
+                "birth": [1, 1, 2, 2, 3, 3, 1, 1, 2, 2, 3, 3, 1, 1, 2, 2, 3, 3],
+                "age": [1, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1, 2, 1, 2],
+            }
+        )
+        expected = expected.set_index(["famid", "birth", "age"])[["ht"]]
+        result = wide_to_long(df, "ht", i=["famid", "birth"], j="age")
+        tm.assert_frame_equal(result, expected)
+
+    def test_non_unique_idvars(self):
+        # GH16382
+        # Raise an error message if non unique id vars (i) are passed
+        df = pd.DataFrame(
+            {"A_A1": [1, 2, 3, 4, 5], "B_B1": [1, 2, 3, 4, 5], "x": [1, 1, 1, 1, 1]}
+        )
+        msg = "the id variables need to uniquely identify each row"
+        with pytest.raises(ValueError, match=msg):
+            wide_to_long(df, ["A_A", "B_B"], i="x", j="colname")
+
+    def test_cast_j_int(self):
+        df = pd.DataFrame(
+            {
+                "actor_1": ["CCH Pounder", "Johnny Depp", "Christoph Waltz"],
+                "actor_2": ["Joel David Moore", "Orlando Bloom", "Rory Kinnear"],
+                "actor_fb_likes_1": [1000.0, 40000.0, 11000.0],
+                "actor_fb_likes_2": [936.0, 5000.0, 393.0],
+                "title": ["Avatar", "Pirates of the Caribbean", "Spectre"],
+            }
+        )
+
+        expected = pd.DataFrame(
+            {
+                "actor": [
+                    "CCH Pounder",
+                    "Johnny Depp",
+                    "Christoph Waltz",
+                    "Joel David Moore",
+                    "Orlando Bloom",
+                    "Rory Kinnear",
+                ],
+                "actor_fb_likes": [1000.0, 40000.0, 11000.0, 936.0, 5000.0, 393.0],
+                "num": [1, 1, 1, 2, 2, 2],
+                "title": [
+                    "Avatar",
+                    "Pirates of the Caribbean",
+                    "Spectre",
+                    "Avatar",
+                    "Pirates of the Caribbean",
+                    "Spectre",
+                ],
+            }
+        ).set_index(["title", "num"])
+        result = wide_to_long(
+            df, ["actor", "actor_fb_likes"], i="title", j="num", sep="_"
+        )
+
+        tm.assert_frame_equal(result, expected)
+
+    def test_identical_stubnames(self):
+        df = pd.DataFrame(
+            {
+                "A2010": [1.0, 2.0],
+                "A2011": [3.0, 4.0],
+                "B2010": [5.0, 6.0],
+                "A": ["X1", "X2"],
+            }
+        )
+        msg = "stubname can't be identical to a column name"
+        with pytest.raises(ValueError, match=msg):
+            wide_to_long(df, ["A", "B"], i="A", j="colname")
+
+    def test_nonnumeric_suffix(self):
+        df = pd.DataFrame(
+            {
+                "treatment_placebo": [1.0, 2.0],
+                "treatment_test": [3.0, 4.0],
+                "result_placebo": [5.0, 6.0],
+                "A": ["X1", "X2"],
+            }
+        )
+        expected = pd.DataFrame(
+            {
+                "A": ["X1", "X1", "X2", "X2"],
+                "colname": ["placebo", "test", "placebo", "test"],
+                "result": [5.0, np.nan, 6.0, np.nan],
+                "treatment": [1.0, 3.0, 2.0, 4.0],
+            }
+        )
+        expected = expected.set_index(["A", "colname"])
+        result = wide_to_long(
+            df, ["result", "treatment"], i="A", j="colname", suffix="[a-z]+", sep="_"
+        )
+        tm.assert_frame_equal(result, expected)
+
+    def test_mixed_type_suffix(self):
+        df = pd.DataFrame(
+            {
+                "A": ["X1", "X2"],
+                "result_1": [0, 9],
+                "result_foo": [5.0, 6.0],
+                "treatment_1": [1.0, 2.0],
+                "treatment_foo": [3.0, 4.0],
+            }
+        )
+        expected = pd.DataFrame(
+            {
+                "A": ["X1", "X2", "X1", "X2"],
+                "colname": ["1", "1", "foo", "foo"],
+                "result": [0.0, 9.0, 5.0, 6.0],
+                "treatment": [1.0, 2.0, 3.0, 4.0],
+            }
+        ).set_index(["A", "colname"])
+        result = wide_to_long(
+            df, ["result", "treatment"], i="A", j="colname", suffix=".+", sep="_"
+        )
+        tm.assert_frame_equal(result, expected)
+
+    def test_float_suffix(self):
+        df = pd.DataFrame(
+            {
+                "treatment_1.1": [1.0, 2.0],
+                "treatment_2.1": [3.0, 4.0],
+                "result_1.2": [5.0, 6.0],
+                "result_1": [0, 9],
+                "A": ["X1", "X2"],
+            }
+        )
+        expected = pd.DataFrame(
+            {
+                "A": ["X1", "X1", "X1", "X1", "X2", "X2", "X2", "X2"],
+                "colname": [1, 1.1, 1.2, 2.1, 1, 1.1, 1.2, 2.1],
+                "result": [0.0, np.nan, 5.0, np.nan, 9.0, np.nan, 6.0, np.nan],
+                "treatment": [np.nan, 1.0, np.nan, 3.0, np.nan, 2.0, np.nan, 4.0],
+            }
+        )
+        expected = expected.set_index(["A", "colname"])
+        result = wide_to_long(
+            df, ["result", "treatment"], i="A", j="colname", suffix="[0-9.]+", sep="_"
+        )
+        tm.assert_frame_equal(result, expected)
+
+    def test_col_substring_of_stubname(self):
+        # GH22468
+        # Don't raise ValueError when a column name is a substring
+        # of a stubname that's been passed as a string
+        wide_data = {
+            "node_id": {0: 0, 1: 1, 2: 2, 3: 3, 4: 4},
+            "A": {0: 0.80, 1: 0.0, 2: 0.25, 3: 1.0, 4: 0.81},
+            "PA0": {0: 0.74, 1: 0.56, 2: 0.56, 3: 0.98, 4: 0.6},
+            "PA1": {0: 0.77, 1: 0.64, 2: 0.52, 3: 0.98, 4: 0.67},
+            "PA3": {0: 0.34, 1: 0.70, 2: 0.52, 3: 0.98, 4: 0.67},
+        }
+        wide_df = pd.DataFrame.from_dict(wide_data)
+        expected = pd.wide_to_long(
+            wide_df, stubnames=["PA"], i=["node_id", "A"], j="time"
+        )
+        result = pd.wide_to_long(wide_df, stubnames="PA", i=["node_id", "A"], j="time")
+        tm.assert_frame_equal(result, expected)

venv/lib/python3.6/site-packages/pandas/tests/reshape/test_qcut.py

@@ -0,0 +1,238 @@
+import os
+
+import numpy as np
+import pytest
+
+from pandas import (
+    Categorical,
+    DatetimeIndex,
+    Interval,
+    IntervalIndex,
+    NaT,
+    Series,
+    TimedeltaIndex,
+    Timestamp,
+    cut,
+    date_range,
+    isna,
+    qcut,
+    timedelta_range,
+)
+from pandas.api.types import CategoricalDtype as CDT
+from pandas.core.algorithms import quantile
+import pandas.util.testing as tm
+
+from pandas.tseries.offsets import Day, Nano
+
+
+def test_qcut():
+    arr = np.random.randn(1000)
+
+    # We store the bins as Index that have been
+    # rounded to comparisons are a bit tricky.
+    labels, bins = qcut(arr, 4, retbins=True)
+    ex_bins = quantile(arr, [0, 0.25, 0.5, 0.75, 1.0])
+
+    result = labels.categories.left.values
+    assert np.allclose(result, ex_bins[:-1], atol=1e-2)
+
+    result = labels.categories.right.values
+    assert np.allclose(result, ex_bins[1:], atol=1e-2)
+
+    ex_levels = cut(arr, ex_bins, include_lowest=True)
+    tm.assert_categorical_equal(labels, ex_levels)
+
+
+def test_qcut_bounds():
+    arr = np.random.randn(1000)
+
+    factor = qcut(arr, 10, labels=False)
+    assert len(np.unique(factor)) == 10
+
+
+def test_qcut_specify_quantiles():
+    arr = np.random.randn(100)
+    factor = qcut(arr, [0, 0.25, 0.5, 0.75, 1.0])
+
+    expected = qcut(arr, 4)
+    tm.assert_categorical_equal(factor, expected)
+
+
+def test_qcut_all_bins_same():
+    with pytest.raises(ValueError, match="edges.*unique"):
+        qcut([0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 3)
+
+
+def test_qcut_include_lowest():
+    values = np.arange(10)
+    ii = qcut(values, 4)
+
+    ex_levels = IntervalIndex(
+        [
+            Interval(-0.001, 2.25),
+            Interval(2.25, 4.5),
+            Interval(4.5, 6.75),
+            Interval(6.75, 9),
+        ]
+    )
+    tm.assert_index_equal(ii.categories, ex_levels)
+
+
+def test_qcut_nas():
+    arr = np.random.randn(100)
+    arr[:20] = np.nan
+
+    result = qcut(arr, 4)
+    assert isna(result[:20]).all()
+
+
+def test_qcut_index():
+    result = qcut([0, 2], 2)
+    intervals = [Interval(-0.001, 1), Interval(1, 2)]
+
+    expected = Categorical(intervals, ordered=True)
+    tm.assert_categorical_equal(result, expected)
+
+
+def test_qcut_binning_issues(datapath):
+    # see gh-1978, gh-1979
+    cut_file = datapath(os.path.join("reshape", "data", "cut_data.csv"))
+    arr = np.loadtxt(cut_file)
+    result = qcut(arr, 20)
+
+    starts = []
+    ends = []
+
+    for lev in np.unique(result):
+        s = lev.left
+        e = lev.right
+        assert s != e
+
+        starts.append(float(s))
+        ends.append(float(e))
+
+    for (sp, sn), (ep, en) in zip(
+        zip(starts[:-1], starts[1:]), zip(ends[:-1], ends[1:])
+    ):
+        assert sp < sn
+        assert ep < en
+        assert ep <= sn
+
+
+def test_qcut_return_intervals():
+    ser = Series([0, 1, 2, 3, 4, 5, 6, 7, 8])
+    res = qcut(ser, [0, 0.333, 0.666, 1])
+
+    exp_levels = np.array(
+        [Interval(-0.001, 2.664), Interval(2.664, 5.328), Interval(5.328, 8)]
+    )
+    exp = Series(exp_levels.take([0, 0, 0, 1, 1, 1, 2, 2, 2])).astype(CDT(ordered=True))
+    tm.assert_series_equal(res, exp)
+
+
+@pytest.mark.parametrize(
+    "kwargs,msg",
+    [
+        (dict(duplicates="drop"), None),
+        (dict(), "Bin edges must be unique"),
+        (dict(duplicates="raise"), "Bin edges must be unique"),
+        (dict(duplicates="foo"), "invalid value for 'duplicates' parameter"),
+    ],
+)
+def test_qcut_duplicates_bin(kwargs, msg):
+    # see gh-7751
+    values = [0, 0, 0, 0, 1, 2, 3]
+
+    if msg is not None:
+        with pytest.raises(ValueError, match=msg):
+            qcut(values, 3, **kwargs)
+    else:
+        result = qcut(values, 3, **kwargs)
+        expected = IntervalIndex([Interval(-0.001, 1), Interval(1, 3)])
+        tm.assert_index_equal(result.categories, expected)
+
+
+@pytest.mark.parametrize(
+    "data,start,end", [(9.0, 8.999, 9.0), (0.0, -0.001, 0.0), (-9.0, -9.001, -9.0)]
+)
+@pytest.mark.parametrize("length", [1, 2])
+@pytest.mark.parametrize("labels", [None, False])
+def test_single_quantile(data, start, end, length, labels):
+    # see gh-15431
+    ser = Series([data] * length)
+    result = qcut(ser, 1, labels=labels)
+
+    if labels is None:
+        intervals = IntervalIndex([Interval(start, end)] * length, closed="right")
+        expected = Series(intervals).astype(CDT(ordered=True))
+    else:
+        expected = Series([0] * length)
+
+    tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "ser",
+    [
+        Series(DatetimeIndex(["20180101", NaT, "20180103"])),
+        Series(TimedeltaIndex(["0 days", NaT, "2 days"])),
+    ],
+    ids=lambda x: str(x.dtype),
+)
+def test_qcut_nat(ser):
+    # see gh-19768
+    intervals = IntervalIndex.from_tuples(
+        [(ser[0] - Nano(), ser[2] - Day()), np.nan, (ser[2] - Day(), ser[2])]
+    )
+    expected = Series(Categorical(intervals, ordered=True))
+
+    result = qcut(ser, 2)
+    tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize("bins", [3, np.linspace(0, 1, 4)])
+def test_datetime_tz_qcut(bins):
+    # see gh-19872
+    tz = "US/Eastern"
+    ser = Series(date_range("20130101", periods=3, tz=tz))
+
+    result = qcut(ser, bins)
+    expected = Series(
+        IntervalIndex(
+            [
+                Interval(
+                    Timestamp("2012-12-31 23:59:59.999999999", tz=tz),
+                    Timestamp("2013-01-01 16:00:00", tz=tz),
+                ),
+                Interval(
+                    Timestamp("2013-01-01 16:00:00", tz=tz),
+                    Timestamp("2013-01-02 08:00:00", tz=tz),
+                ),
+                Interval(
+                    Timestamp("2013-01-02 08:00:00", tz=tz),
+                    Timestamp("2013-01-03 00:00:00", tz=tz),
+                ),
+            ]
+        )
+    ).astype(CDT(ordered=True))
+    tm.assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "arg,expected_bins",
+    [
+        [
+            timedelta_range("1day", periods=3),
+            TimedeltaIndex(["1 days", "2 days", "3 days"]),
+        ],
+        [
+            date_range("20180101", periods=3),
+            DatetimeIndex(["2018-01-01", "2018-01-02", "2018-01-03"]),
+        ],
+    ],
+)
+def test_date_like_qcut_bins(arg, expected_bins):
+    # see gh-19891
+    ser = Series(arg)
+    result, result_bins = qcut(ser, 2, retbins=True)
+    tm.assert_index_equal(result_bins, expected_bins)

venv/lib/python3.6/site-packages/pandas/tests/reshape/test_reshape.py

@@ -0,0 +1,653 @@
+from collections import OrderedDict
+
+import numpy as np
+from numpy import nan
+import pytest
+
+from pandas.core.dtypes.common import is_integer_dtype
+
+import pandas as pd
+from pandas import Categorical, DataFrame, Index, Series, get_dummies
+from pandas.core.sparse.api import SparseArray, SparseDtype
+import pandas.util.testing as tm
+from pandas.util.testing import assert_frame_equal
+
+
+@pytest.mark.filterwarnings("ignore:Sparse:FutureWarning")
+class TestGetDummies:
+    @pytest.fixture
+    def df(self):
+        return DataFrame({"A": ["a", "b", "a"], "B": ["b", "b", "c"], "C": [1, 2, 3]})
+
+    @pytest.fixture(params=["uint8", "i8", np.float64, bool, None])
+    def dtype(self, request):
+        return np.dtype(request.param)
+
+    @pytest.fixture(params=["dense", "sparse"])
+    def sparse(self, request):
+        # params are strings to simplify reading test results,
+        # e.g. TestGetDummies::test_basic[uint8-sparse] instead of [uint8-True]
+        return request.param == "sparse"
+
+    def effective_dtype(self, dtype):
+        if dtype is None:
+            return np.uint8
+        return dtype
+
+    def test_raises_on_dtype_object(self, df):
+        with pytest.raises(ValueError):
+            get_dummies(df, dtype="object")
+
+    def test_basic(self, sparse, dtype):
+        s_list = list("abc")
+        s_series = Series(s_list)
+        s_series_index = Series(s_list, list("ABC"))
+
+        expected = DataFrame(
+            {"a": [1, 0, 0], "b": [0, 1, 0], "c": [0, 0, 1]},
+            dtype=self.effective_dtype(dtype),
+        )
+        if sparse:
+            expected = expected.apply(pd.SparseArray, fill_value=0.0)
+        result = get_dummies(s_list, sparse=sparse, dtype=dtype)
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(s_series, sparse=sparse, dtype=dtype)
+        assert_frame_equal(result, expected)
+
+        expected.index = list("ABC")
+        result = get_dummies(s_series_index, sparse=sparse, dtype=dtype)
+        assert_frame_equal(result, expected)
+
+    def test_basic_types(self, sparse, dtype):
+        # GH 10531
+        s_list = list("abc")
+        s_series = Series(s_list)
+        s_df = DataFrame(
+            {"a": [0, 1, 0, 1, 2], "b": ["A", "A", "B", "C", "C"], "c": [2, 3, 3, 3, 2]}
+        )
+
+        expected = DataFrame(
+            {"a": [1, 0, 0], "b": [0, 1, 0], "c": [0, 0, 1]},
+            dtype=self.effective_dtype(dtype),
+            columns=list("abc"),
+        )
+        if sparse:
+            if is_integer_dtype(dtype):
+                fill_value = 0
+            elif dtype == bool:
+                fill_value = False
+            else:
+                fill_value = 0.0
+
+            expected = expected.apply(SparseArray, fill_value=fill_value)
+        result = get_dummies(s_list, sparse=sparse, dtype=dtype)
+        tm.assert_frame_equal(result, expected)
+
+        result = get_dummies(s_series, sparse=sparse, dtype=dtype)
+        tm.assert_frame_equal(result, expected)
+
+        result = get_dummies(s_df, columns=s_df.columns, sparse=sparse, dtype=dtype)
+        if sparse:
+            dtype_name = "Sparse[{}, {}]".format(
+                self.effective_dtype(dtype).name, fill_value
+            )
+        else:
+            dtype_name = self.effective_dtype(dtype).name
+
+        expected = Series({dtype_name: 8})
+        result = result.dtypes.value_counts()
+        result.index = [str(i) for i in result.index]
+        tm.assert_series_equal(result, expected)
+
+        result = get_dummies(s_df, columns=["a"], sparse=sparse, dtype=dtype)
+
+        expected_counts = {"int64": 1, "object": 1}
+        expected_counts[dtype_name] = 3 + expected_counts.get(dtype_name, 0)
+
+        expected = Series(expected_counts).sort_index()
+        result = result.dtypes.value_counts()
+        result.index = [str(i) for i in result.index]
+        result = result.sort_index()
+        tm.assert_series_equal(result, expected)
+
+    def test_just_na(self, sparse):
+        just_na_list = [np.nan]
+        just_na_series = Series(just_na_list)
+        just_na_series_index = Series(just_na_list, index=["A"])
+
+        res_list = get_dummies(just_na_list, sparse=sparse)
+        res_series = get_dummies(just_na_series, sparse=sparse)
+        res_series_index = get_dummies(just_na_series_index, sparse=sparse)
+
+        assert res_list.empty
+        assert res_series.empty
+        assert res_series_index.empty
+
+        assert res_list.index.tolist() == [0]
+        assert res_series.index.tolist() == [0]
+        assert res_series_index.index.tolist() == ["A"]
+
+    def test_include_na(self, sparse, dtype):
+        s = ["a", "b", np.nan]
+        res = get_dummies(s, sparse=sparse, dtype=dtype)
+        exp = DataFrame(
+            {"a": [1, 0, 0], "b": [0, 1, 0]}, dtype=self.effective_dtype(dtype)
+        )
+        if sparse:
+            exp = exp.apply(pd.SparseArray, fill_value=0.0)
+        assert_frame_equal(res, exp)
+
+        # Sparse dataframes do not allow nan labelled columns, see #GH8822
+        res_na = get_dummies(s, dummy_na=True, sparse=sparse, dtype=dtype)
+        exp_na = DataFrame(
+            {nan: [0, 0, 1], "a": [1, 0, 0], "b": [0, 1, 0]},
+            dtype=self.effective_dtype(dtype),
+        )
+        exp_na = exp_na.reindex(["a", "b", nan], axis=1)
+        # hack (NaN handling in assert_index_equal)
+        exp_na.columns = res_na.columns
+        if sparse:
+            exp_na = exp_na.apply(pd.SparseArray, fill_value=0.0)
+        assert_frame_equal(res_na, exp_na)
+
+        res_just_na = get_dummies([nan], dummy_na=True, sparse=sparse, dtype=dtype)
+        exp_just_na = DataFrame(
+            Series(1, index=[0]), columns=[nan], dtype=self.effective_dtype(dtype)
+        )
+        tm.assert_numpy_array_equal(res_just_na.values, exp_just_na.values)
+
+    def test_unicode(self, sparse):
+        # See GH 6885 - get_dummies chokes on unicode values
+        import unicodedata
+
+        e = "e"
+        eacute = unicodedata.lookup("LATIN SMALL LETTER E WITH ACUTE")
+        s = [e, eacute, eacute]
+        res = get_dummies(s, prefix="letter", sparse=sparse)
+        exp = DataFrame(
+            {"letter_e": [1, 0, 0], "letter_{eacute}".format(eacute=eacute): [0, 1, 1]},
+            dtype=np.uint8,
+        )
+        if sparse:
+            exp = exp.apply(pd.SparseArray, fill_value=0)
+        assert_frame_equal(res, exp)
+
+    def test_dataframe_dummies_all_obj(self, df, sparse):
+        df = df[["A", "B"]]
+        result = get_dummies(df, sparse=sparse)
+        expected = DataFrame(
+            {"A_a": [1, 0, 1], "A_b": [0, 1, 0], "B_b": [1, 1, 0], "B_c": [0, 0, 1]},
+            dtype=np.uint8,
+        )
+        if sparse:
+            expected = pd.DataFrame(
+                {
+                    "A_a": pd.SparseArray([1, 0, 1], dtype="uint8"),
+                    "A_b": pd.SparseArray([0, 1, 0], dtype="uint8"),
+                    "B_b": pd.SparseArray([1, 1, 0], dtype="uint8"),
+                    "B_c": pd.SparseArray([0, 0, 1], dtype="uint8"),
+                }
+            )
+
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_mix_default(self, df, sparse, dtype):
+        result = get_dummies(df, sparse=sparse, dtype=dtype)
+        if sparse:
+            arr = SparseArray
+            typ = SparseDtype(dtype, 0)
+        else:
+            arr = np.array
+            typ = dtype
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3],
+                "A_a": arr([1, 0, 1], dtype=typ),
+                "A_b": arr([0, 1, 0], dtype=typ),
+                "B_b": arr([1, 1, 0], dtype=typ),
+                "B_c": arr([0, 0, 1], dtype=typ),
+            }
+        )
+        expected = expected[["C", "A_a", "A_b", "B_b", "B_c"]]
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_prefix_list(self, df, sparse):
+        prefixes = ["from_A", "from_B"]
+        result = get_dummies(df, prefix=prefixes, sparse=sparse)
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3],
+                "from_A_a": [1, 0, 1],
+                "from_A_b": [0, 1, 0],
+                "from_B_b": [1, 1, 0],
+                "from_B_c": [0, 0, 1],
+            },
+            dtype=np.uint8,
+        )
+        expected[["C"]] = df[["C"]]
+        cols = ["from_A_a", "from_A_b", "from_B_b", "from_B_c"]
+        expected = expected[["C"] + cols]
+
+        typ = pd.SparseArray if sparse else pd.Series
+        expected[cols] = expected[cols].apply(lambda x: typ(x))
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_prefix_str(self, df, sparse):
+        # not that you should do this...
+        result = get_dummies(df, prefix="bad", sparse=sparse)
+        bad_columns = ["bad_a", "bad_b", "bad_b", "bad_c"]
+        expected = DataFrame(
+            [[1, 1, 0, 1, 0], [2, 0, 1, 1, 0], [3, 1, 0, 0, 1]],
+            columns=["C"] + bad_columns,
+            dtype=np.uint8,
+        )
+        expected = expected.astype({"C": np.int64})
+        if sparse:
+            # work around astyping & assigning with duplicate columns
+            # https://github.com/pandas-dev/pandas/issues/14427
+            expected = pd.concat(
+                [
+                    pd.Series([1, 2, 3], name="C"),
+                    pd.Series([1, 0, 1], name="bad_a", dtype="Sparse[uint8]"),
+                    pd.Series([0, 1, 0], name="bad_b", dtype="Sparse[uint8]"),
+                    pd.Series([1, 1, 0], name="bad_b", dtype="Sparse[uint8]"),
+                    pd.Series([0, 0, 1], name="bad_c", dtype="Sparse[uint8]"),
+                ],
+                axis=1,
+            )
+
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_subset(self, df, sparse):
+        result = get_dummies(df, prefix=["from_A"], columns=["A"], sparse=sparse)
+        expected = DataFrame(
+            {
+                "B": ["b", "b", "c"],
+                "C": [1, 2, 3],
+                "from_A_a": [1, 0, 1],
+                "from_A_b": [0, 1, 0],
+            },
+            dtype=np.uint8,
+        )
+        expected[["C"]] = df[["C"]]
+        if sparse:
+            cols = ["from_A_a", "from_A_b"]
+            expected[cols] = expected[cols].apply(lambda x: pd.SparseSeries(x))
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_prefix_sep(self, df, sparse):
+        result = get_dummies(df, prefix_sep="..", sparse=sparse)
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3],
+                "A..a": [1, 0, 1],
+                "A..b": [0, 1, 0],
+                "B..b": [1, 1, 0],
+                "B..c": [0, 0, 1],
+            },
+            dtype=np.uint8,
+        )
+        expected[["C"]] = df[["C"]]
+        expected = expected[["C", "A..a", "A..b", "B..b", "B..c"]]
+        if sparse:
+            cols = ["A..a", "A..b", "B..b", "B..c"]
+            expected[cols] = expected[cols].apply(lambda x: pd.SparseSeries(x))
+
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(df, prefix_sep=["..", "__"], sparse=sparse)
+        expected = expected.rename(columns={"B..b": "B__b", "B..c": "B__c"})
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(df, prefix_sep={"A": "..", "B": "__"}, sparse=sparse)
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_prefix_bad_length(self, df, sparse):
+        with pytest.raises(ValueError):
+            get_dummies(df, prefix=["too few"], sparse=sparse)
+
+    def test_dataframe_dummies_prefix_sep_bad_length(self, df, sparse):
+        with pytest.raises(ValueError):
+            get_dummies(df, prefix_sep=["bad"], sparse=sparse)
+
+    def test_dataframe_dummies_prefix_dict(self, sparse):
+        prefixes = {"A": "from_A", "B": "from_B"}
+        df = DataFrame({"C": [1, 2, 3], "A": ["a", "b", "a"], "B": ["b", "b", "c"]})
+        result = get_dummies(df, prefix=prefixes, sparse=sparse)
+
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3],
+                "from_A_a": [1, 0, 1],
+                "from_A_b": [0, 1, 0],
+                "from_B_b": [1, 1, 0],
+                "from_B_c": [0, 0, 1],
+            }
+        )
+
+        columns = ["from_A_a", "from_A_b", "from_B_b", "from_B_c"]
+        expected[columns] = expected[columns].astype(np.uint8)
+        if sparse:
+            expected[columns] = expected[columns].apply(lambda x: pd.SparseSeries(x))
+
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_with_na(self, df, sparse, dtype):
+        df.loc[3, :] = [np.nan, np.nan, np.nan]
+        result = get_dummies(df, dummy_na=True, sparse=sparse, dtype=dtype).sort_index(
+            axis=1
+        )
+
+        if sparse:
+            arr = SparseArray
+            typ = SparseDtype(dtype, 0)
+        else:
+            arr = np.array
+            typ = dtype
+
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3, np.nan],
+                "A_a": arr([1, 0, 1, 0], dtype=typ),
+                "A_b": arr([0, 1, 0, 0], dtype=typ),
+                "A_nan": arr([0, 0, 0, 1], dtype=typ),
+                "B_b": arr([1, 1, 0, 0], dtype=typ),
+                "B_c": arr([0, 0, 1, 0], dtype=typ),
+                "B_nan": arr([0, 0, 0, 1], dtype=typ),
+            }
+        ).sort_index(axis=1)
+
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(df, dummy_na=False, sparse=sparse, dtype=dtype)
+        expected = expected[["C", "A_a", "A_b", "B_b", "B_c"]]
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_with_categorical(self, df, sparse, dtype):
+        df["cat"] = pd.Categorical(["x", "y", "y"])
+        result = get_dummies(df, sparse=sparse, dtype=dtype).sort_index(axis=1)
+        if sparse:
+            arr = SparseArray
+            typ = SparseDtype(dtype, 0)
+        else:
+            arr = np.array
+            typ = dtype
+
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3],
+                "A_a": arr([1, 0, 1], dtype=typ),
+                "A_b": arr([0, 1, 0], dtype=typ),
+                "B_b": arr([1, 1, 0], dtype=typ),
+                "B_c": arr([0, 0, 1], dtype=typ),
+                "cat_x": arr([1, 0, 0], dtype=typ),
+                "cat_y": arr([0, 1, 1], dtype=typ),
+            }
+        ).sort_index(axis=1)
+
+        assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize(
+        "get_dummies_kwargs,expected",
+        [
+            (
+                {"data": pd.DataFrame(({"ä": ["a"]}))},
+                pd.DataFrame({"ä_a": [1]}, dtype=np.uint8),
+            ),
+            (
+                {"data": pd.DataFrame({"x": ["ä"]})},
+                pd.DataFrame({"x_ä": [1]}, dtype=np.uint8),
+            ),
+            (
+                {"data": pd.DataFrame({"x": ["a"]}), "prefix": "ä"},
+                pd.DataFrame({"ä_a": [1]}, dtype=np.uint8),
+            ),
+            (
+                {"data": pd.DataFrame({"x": ["a"]}), "prefix_sep": "ä"},
+                pd.DataFrame({"xäa": [1]}, dtype=np.uint8),
+            ),
+        ],
+    )
+    def test_dataframe_dummies_unicode(self, get_dummies_kwargs, expected):
+        # GH22084 pd.get_dummies incorrectly encodes unicode characters
+        # in dataframe column names
+        result = get_dummies(**get_dummies_kwargs)
+        assert_frame_equal(result, expected)
+
+    def test_basic_drop_first(self, sparse):
+        # GH12402 Add a new parameter `drop_first` to avoid collinearity
+        # Basic case
+        s_list = list("abc")
+        s_series = Series(s_list)
+        s_series_index = Series(s_list, list("ABC"))
+
+        expected = DataFrame({"b": [0, 1, 0], "c": [0, 0, 1]}, dtype=np.uint8)
+
+        result = get_dummies(s_list, drop_first=True, sparse=sparse)
+        if sparse:
+            expected = expected.apply(pd.SparseArray, fill_value=0)
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(s_series, drop_first=True, sparse=sparse)
+        assert_frame_equal(result, expected)
+
+        expected.index = list("ABC")
+        result = get_dummies(s_series_index, drop_first=True, sparse=sparse)
+        assert_frame_equal(result, expected)
+
+    def test_basic_drop_first_one_level(self, sparse):
+        # Test the case that categorical variable only has one level.
+        s_list = list("aaa")
+        s_series = Series(s_list)
+        s_series_index = Series(s_list, list("ABC"))
+
+        expected = DataFrame(index=np.arange(3))
+
+        result = get_dummies(s_list, drop_first=True, sparse=sparse)
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(s_series, drop_first=True, sparse=sparse)
+        assert_frame_equal(result, expected)
+
+        expected = DataFrame(index=list("ABC"))
+        result = get_dummies(s_series_index, drop_first=True, sparse=sparse)
+        assert_frame_equal(result, expected)
+
+    def test_basic_drop_first_NA(self, sparse):
+        # Test NA handling together with drop_first
+        s_NA = ["a", "b", np.nan]
+        res = get_dummies(s_NA, drop_first=True, sparse=sparse)
+        exp = DataFrame({"b": [0, 1, 0]}, dtype=np.uint8)
+        if sparse:
+            exp = exp.apply(pd.SparseArray, fill_value=0)
+
+        assert_frame_equal(res, exp)
+
+        res_na = get_dummies(s_NA, dummy_na=True, drop_first=True, sparse=sparse)
+        exp_na = DataFrame({"b": [0, 1, 0], nan: [0, 0, 1]}, dtype=np.uint8).reindex(
+            ["b", nan], axis=1
+        )
+        if sparse:
+            exp_na = exp_na.apply(pd.SparseArray, fill_value=0)
+        assert_frame_equal(res_na, exp_na)
+
+        res_just_na = get_dummies([nan], dummy_na=True, drop_first=True, sparse=sparse)
+        exp_just_na = DataFrame(index=np.arange(1))
+        assert_frame_equal(res_just_na, exp_just_na)
+
+    def test_dataframe_dummies_drop_first(self, df, sparse):
+        df = df[["A", "B"]]
+        result = get_dummies(df, drop_first=True, sparse=sparse)
+        expected = DataFrame({"A_b": [0, 1, 0], "B_c": [0, 0, 1]}, dtype=np.uint8)
+        if sparse:
+            expected = expected.apply(pd.SparseArray, fill_value=0)
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_drop_first_with_categorical(self, df, sparse, dtype):
+        df["cat"] = pd.Categorical(["x", "y", "y"])
+        result = get_dummies(df, drop_first=True, sparse=sparse)
+        expected = DataFrame(
+            {"C": [1, 2, 3], "A_b": [0, 1, 0], "B_c": [0, 0, 1], "cat_y": [0, 1, 1]}
+        )
+        cols = ["A_b", "B_c", "cat_y"]
+        expected[cols] = expected[cols].astype(np.uint8)
+        expected = expected[["C", "A_b", "B_c", "cat_y"]]
+        if sparse:
+            for col in cols:
+                expected[col] = pd.SparseSeries(expected[col])
+        assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_drop_first_with_na(self, df, sparse):
+        df.loc[3, :] = [np.nan, np.nan, np.nan]
+        result = get_dummies(
+            df, dummy_na=True, drop_first=True, sparse=sparse
+        ).sort_index(axis=1)
+        expected = DataFrame(
+            {
+                "C": [1, 2, 3, np.nan],
+                "A_b": [0, 1, 0, 0],
+                "A_nan": [0, 0, 0, 1],
+                "B_c": [0, 0, 1, 0],
+                "B_nan": [0, 0, 0, 1],
+            }
+        )
+        cols = ["A_b", "A_nan", "B_c", "B_nan"]
+        expected[cols] = expected[cols].astype(np.uint8)
+        expected = expected.sort_index(axis=1)
+        if sparse:
+            for col in cols:
+                expected[col] = pd.SparseSeries(expected[col])
+
+        assert_frame_equal(result, expected)
+
+        result = get_dummies(df, dummy_na=False, drop_first=True, sparse=sparse)
+        expected = expected[["C", "A_b", "B_c"]]
+        assert_frame_equal(result, expected)
+
+    def test_int_int(self):
+        data = Series([1, 2, 1])
+        result = pd.get_dummies(data)
+        expected = DataFrame([[1, 0], [0, 1], [1, 0]], columns=[1, 2], dtype=np.uint8)
+        tm.assert_frame_equal(result, expected)
+
+        data = Series(pd.Categorical(["a", "b", "a"]))
+        result = pd.get_dummies(data)
+        expected = DataFrame(
+            [[1, 0], [0, 1], [1, 0]], columns=pd.Categorical(["a", "b"]), dtype=np.uint8
+        )
+        tm.assert_frame_equal(result, expected)
+
+    def test_int_df(self, dtype):
+        data = DataFrame(
+            {
+                "A": [1, 2, 1],
+                "B": pd.Categorical(["a", "b", "a"]),
+                "C": [1, 2, 1],
+                "D": [1.0, 2.0, 1.0],
+            }
+        )
+        columns = ["C", "D", "A_1", "A_2", "B_a", "B_b"]
+        expected = DataFrame(
+            [[1, 1.0, 1, 0, 1, 0], [2, 2.0, 0, 1, 0, 1], [1, 1.0, 1, 0, 1, 0]],
+            columns=columns,
+        )
+        expected[columns[2:]] = expected[columns[2:]].astype(dtype)
+        result = pd.get_dummies(data, columns=["A", "B"], dtype=dtype)
+        tm.assert_frame_equal(result, expected)
+
+    def test_dataframe_dummies_preserve_categorical_dtype(self, dtype):
+        # GH13854
+        for ordered in [False, True]:
+            cat = pd.Categorical(list("xy"), categories=list("xyz"), ordered=ordered)
+            result = get_dummies(cat, dtype=dtype)
+
+            data = np.array([[1, 0, 0], [0, 1, 0]], dtype=self.effective_dtype(dtype))
+            cols = pd.CategoricalIndex(
+                cat.categories, categories=cat.categories, ordered=ordered
+            )
+            expected = DataFrame(data, columns=cols, dtype=self.effective_dtype(dtype))
+
+            tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("sparse", [True, False])
+    def test_get_dummies_dont_sparsify_all_columns(self, sparse):
+        # GH18914
+        df = DataFrame.from_dict(
+            OrderedDict([("GDP", [1, 2]), ("Nation", ["AB", "CD"])])
+        )
+        df = get_dummies(df, columns=["Nation"], sparse=sparse)
+        df2 = df.reindex(columns=["GDP"])
+
+        tm.assert_frame_equal(df[["GDP"]], df2)
+
+    def test_get_dummies_duplicate_columns(self, df):
+        # GH20839
+        df.columns = ["A", "A", "A"]
+        result = get_dummies(df).sort_index(axis=1)
+
+        expected = DataFrame(
+            [[1, 1, 0, 1, 0], [2, 0, 1, 1, 0], [3, 1, 0, 0, 1]],
+            columns=["A", "A_a", "A_b", "A_b", "A_c"],
+            dtype=np.uint8,
+        ).sort_index(axis=1)
+
+        expected = expected.astype({"A": np.int64})
+
+        tm.assert_frame_equal(result, expected)
+
+    def test_get_dummies_all_sparse(self):
+        df = pd.DataFrame({"A": [1, 2]})
+        result = pd.get_dummies(df, columns=["A"], sparse=True)
+        dtype = SparseDtype("uint8", 0)
+        expected = pd.DataFrame(
+            {
+                "A_1": SparseArray([1, 0], dtype=dtype),
+                "A_2": SparseArray([0, 1], dtype=dtype),
+            }
+        )
+        tm.assert_frame_equal(result, expected)
+
+
+class TestCategoricalReshape:
+    def test_reshaping_multi_index_categorical(self):
+
+        cols = ["ItemA", "ItemB", "ItemC"]
+        data = {c: tm.makeTimeDataFrame() for c in cols}
+        df = pd.concat({c: data[c].stack() for c in data}, axis="columns")
+        df.index.names = ["major", "minor"]
+        df["str"] = "foo"
+
+        dti = df.index.levels[0]
+
+        df["category"] = df["str"].astype("category")
+        result = df["category"].unstack()
+
+        c = Categorical(["foo"] * len(dti))
+        expected = DataFrame(
+            {"A": c.copy(), "B": c.copy(), "C": c.copy(), "D": c.copy()},
+            columns=Index(list("ABCD"), name="minor"),
+            index=dti,
+        )
+        tm.assert_frame_equal(result, expected)
+
+
+class TestMakeAxisDummies:
+    def test_preserve_categorical_dtype(self):
+        # GH13854
+        for ordered in [False, True]:
+            cidx = pd.CategoricalIndex(list("xyz"), ordered=ordered)
+            midx = pd.MultiIndex(levels=[["a"], cidx], codes=[[0, 0], [0, 1]])
+            df = DataFrame([[10, 11]], index=midx)
+
+            expected = DataFrame(
+                [[1.0, 0.0, 0.0], [0.0, 1.0, 0.0]], index=midx, columns=cidx
+            )
+
+            from pandas.core.reshape.reshape import make_axis_dummies
+
+            result = make_axis_dummies(df)
+            tm.assert_frame_equal(result, expected)
+
+            result = make_axis_dummies(df, transform=lambda x: x)
+            tm.assert_frame_equal(result, expected)

venv/lib/python3.6/site-packages/pandas/tests/reshape/test_union_categoricals.py

@@ -0,0 +1,348 @@
+import numpy as np
+import pytest
+
+from pandas.core.dtypes.concat import union_categoricals
+
+import pandas as pd
+from pandas import Categorical, CategoricalIndex, Series
+from pandas.util import testing as tm
+
+
+class TestUnionCategoricals:
+    def test_union_categorical(self):
+        # GH 13361
+        data = [
+            (list("abc"), list("abd"), list("abcabd")),
+            ([0, 1, 2], [2, 3, 4], [0, 1, 2, 2, 3, 4]),
+            ([0, 1.2, 2], [2, 3.4, 4], [0, 1.2, 2, 2, 3.4, 4]),
+            (
+                ["b", "b", np.nan, "a"],
+                ["a", np.nan, "c"],
+                ["b", "b", np.nan, "a", "a", np.nan, "c"],
+            ),
+            (
+                pd.date_range("2014-01-01", "2014-01-05"),
+                pd.date_range("2014-01-06", "2014-01-07"),
+                pd.date_range("2014-01-01", "2014-01-07"),
+            ),
+            (
+                pd.date_range("2014-01-01", "2014-01-05", tz="US/Central"),
+                pd.date_range("2014-01-06", "2014-01-07", tz="US/Central"),
+                pd.date_range("2014-01-01", "2014-01-07", tz="US/Central"),
+            ),
+            (
+                pd.period_range("2014-01-01", "2014-01-05"),
+                pd.period_range("2014-01-06", "2014-01-07"),
+                pd.period_range("2014-01-01", "2014-01-07"),
+            ),
+        ]
+
+        for a, b, combined in data:
+            for box in [Categorical, CategoricalIndex, Series]:
+                result = union_categoricals([box(Categorical(a)), box(Categorical(b))])
+                expected = Categorical(combined)
+                tm.assert_categorical_equal(result, expected, check_category_order=True)
+
+        # new categories ordered by appearance
+        s = Categorical(["x", "y", "z"])
+        s2 = Categorical(["a", "b", "c"])
+        result = union_categoricals([s, s2])
+        expected = Categorical(
+            ["x", "y", "z", "a", "b", "c"], categories=["x", "y", "z", "a", "b", "c"]
+        )
+        tm.assert_categorical_equal(result, expected)
+
+        s = Categorical([0, 1.2, 2], ordered=True)
+        s2 = Categorical([0, 1.2, 2], ordered=True)
+        result = union_categoricals([s, s2])
+        expected = Categorical([0, 1.2, 2, 0, 1.2, 2], ordered=True)
+        tm.assert_categorical_equal(result, expected)
+
+        # must exactly match types
+        s = Categorical([0, 1.2, 2])
+        s2 = Categorical([2, 3, 4])
+        msg = "dtype of categories must be the same"
+        with pytest.raises(TypeError, match=msg):
+            union_categoricals([s, s2])
+
+        msg = "No Categoricals to union"
+        with pytest.raises(ValueError, match=msg):
+            union_categoricals([])
+
+    def test_union_categoricals_nan(self):
+        # GH 13759
+        res = union_categoricals(
+            [pd.Categorical([1, 2, np.nan]), pd.Categorical([3, 2, np.nan])]
+        )
+        exp = Categorical([1, 2, np.nan, 3, 2, np.nan])
+        tm.assert_categorical_equal(res, exp)
+
+        res = union_categoricals(
+            [pd.Categorical(["A", "B"]), pd.Categorical(["B", "B", np.nan])]
+        )
+        exp = Categorical(["A", "B", "B", "B", np.nan])
+        tm.assert_categorical_equal(res, exp)
+
+        val1 = [pd.Timestamp("2011-01-01"), pd.Timestamp("2011-03-01"), pd.NaT]
+        val2 = [pd.NaT, pd.Timestamp("2011-01-01"), pd.Timestamp("2011-02-01")]
+
+        res = union_categoricals([pd.Categorical(val1), pd.Categorical(val2)])
+        exp = Categorical(
+            val1 + val2,
+            categories=[
+                pd.Timestamp("2011-01-01"),
+                pd.Timestamp("2011-03-01"),
+                pd.Timestamp("2011-02-01"),
+            ],
+        )
+        tm.assert_categorical_equal(res, exp)
+
+        # all NaN
+        res = union_categoricals(
+            [
+                pd.Categorical(np.array([np.nan, np.nan], dtype=object)),
+                pd.Categorical(["X"]),
+            ]
+        )
+        exp = Categorical([np.nan, np.nan, "X"])
+        tm.assert_categorical_equal(res, exp)
+
+        res = union_categoricals(
+            [pd.Categorical([np.nan, np.nan]), pd.Categorical([np.nan, np.nan])]
+        )
+        exp = Categorical([np.nan, np.nan, np.nan, np.nan])
+        tm.assert_categorical_equal(res, exp)
+
+    def test_union_categoricals_empty(self):
+        # GH 13759
+        res = union_categoricals([pd.Categorical([]), pd.Categorical([])])
+        exp = Categorical([])
+        tm.assert_categorical_equal(res, exp)
+
+        res = union_categoricals([Categorical([]), Categorical(["1"])])
+        exp = Categorical(["1"])
+        tm.assert_categorical_equal(res, exp)
+
+    def test_union_categorical_same_category(self):
+        # check fastpath
+        c1 = Categorical([1, 2, 3, 4], categories=[1, 2, 3, 4])
+        c2 = Categorical([3, 2, 1, np.nan], categories=[1, 2, 3, 4])
+        res = union_categoricals([c1, c2])
+        exp = Categorical([1, 2, 3, 4, 3, 2, 1, np.nan], categories=[1, 2, 3, 4])
+        tm.assert_categorical_equal(res, exp)
+
+        c1 = Categorical(["z", "z", "z"], categories=["x", "y", "z"])
+        c2 = Categorical(["x", "x", "x"], categories=["x", "y", "z"])
+        res = union_categoricals([c1, c2])
+        exp = Categorical(["z", "z", "z", "x", "x", "x"], categories=["x", "y", "z"])
+        tm.assert_categorical_equal(res, exp)
+
+    def test_union_categorical_same_categories_different_order(self):
+        # https://github.com/pandas-dev/pandas/issues/19096
+        c1 = Categorical(["a", "b", "c"], categories=["a", "b", "c"])
+        c2 = Categorical(["a", "b", "c"], categories=["b", "a", "c"])
+        result = union_categoricals([c1, c2])
+        expected = Categorical(
+            ["a", "b", "c", "a", "b", "c"], categories=["a", "b", "c"]
+        )
+        tm.assert_categorical_equal(result, expected)
+
+    def test_union_categoricals_ordered(self):
+        c1 = Categorical([1, 2, 3], ordered=True)
+        c2 = Categorical([1, 2, 3], ordered=False)
+
+        msg = "Categorical.ordered must be the same"
+        with pytest.raises(TypeError, match=msg):
+            union_categoricals([c1, c2])
+
+        res = union_categoricals([c1, c1])
+        exp = Categorical([1, 2, 3, 1, 2, 3], ordered=True)
+        tm.assert_categorical_equal(res, exp)
+
+        c1 = Categorical([1, 2, 3, np.nan], ordered=True)
+        c2 = Categorical([3, 2], categories=[1, 2, 3], ordered=True)
+
+        res = union_categoricals([c1, c2])
+        exp = Categorical([1, 2, 3, np.nan, 3, 2], ordered=True)
+        tm.assert_categorical_equal(res, exp)
+
+        c1 = Categorical([1, 2, 3], ordered=True)
+        c2 = Categorical([1, 2, 3], categories=[3, 2, 1], ordered=True)
+
+        msg = "to union ordered Categoricals, all categories must be the same"
+        with pytest.raises(TypeError, match=msg):
+            union_categoricals([c1, c2])
+
+    def test_union_categoricals_ignore_order(self):
+        # GH 15219
+        c1 = Categorical([1, 2, 3], ordered=True)
+        c2 = Categorical([1, 2, 3], ordered=False)
+
+        res = union_categoricals([c1, c2], ignore_order=True)
+        exp = Categorical([1, 2, 3, 1, 2, 3])
+        tm.assert_categorical_equal(res, exp)
+
+        msg = "Categorical.ordered must be the same"
+        with pytest.raises(TypeError, match=msg):
+            union_categoricals([c1, c2], ignore_order=False)
+
+        res = union_categoricals([c1, c1], ignore_order=True)
+        exp = Categorical([1, 2, 3, 1, 2, 3])
+        tm.assert_categorical_equal(res, exp)
+
+        res = union_categoricals([c1, c1], ignore_order=False)
+        exp = Categorical([1, 2, 3, 1, 2, 3], categories=[1, 2, 3], ordered=True)
+        tm.assert_categorical_equal(res, exp)
+
+        c1 = Categorical([1, 2, 3, np.nan], ordered=True)
+        c2 = Categorical([3, 2], categories=[1, 2, 3], ordered=True)
+
+        res = union_categoricals([c1, c2], ignore_order=True)
+        exp = Categorical([1, 2, 3, np.nan, 3, 2])
+        tm.assert_categorical_equal(res, exp)
+
+        c1 = Categorical([1, 2, 3], ordered=True)
+        c2 = Categorical([1, 2, 3], categories=[3, 2, 1], ordered=True)
+
+        res = union_categoricals([c1, c2], ignore_order=True)
+        exp = Categorical([1, 2, 3, 1, 2, 3])
+        tm.assert_categorical_equal(res, exp)
+
+        res = union_categoricals([c2, c1], ignore_order=True, sort_categories=True)
+        exp = Categorical([1, 2, 3, 1, 2, 3], categories=[1, 2, 3])
+        tm.assert_categorical_equal(res, exp)
+
+        c1 = Categorical([1, 2, 3], ordered=True)
+        c2 = Categorical([4, 5, 6], ordered=True)
+        result = union_categoricals([c1, c2], ignore_order=True)
+        expected = Categorical([1, 2, 3, 4, 5, 6])
+        tm.assert_categorical_equal(result, expected)
+
+        msg = "to union ordered Categoricals, all categories must be the same"
+        with pytest.raises(TypeError, match=msg):
+            union_categoricals([c1, c2], ignore_order=False)
+
+        with pytest.raises(TypeError, match=msg):
+            union_categoricals([c1, c2])
+
+    def test_union_categoricals_sort(self):
+        # GH 13846
+        c1 = Categorical(["x", "y", "z"])
+        c2 = Categorical(["a", "b", "c"])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical(
+            ["x", "y", "z", "a", "b", "c"], categories=["a", "b", "c", "x", "y", "z"]
+        )
+        tm.assert_categorical_equal(result, expected)
+
+        # fastpath
+        c1 = Categorical(["a", "b"], categories=["b", "a", "c"])
+        c2 = Categorical(["b", "c"], categories=["b", "a", "c"])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical(["a", "b", "b", "c"], categories=["a", "b", "c"])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical(["a", "b"], categories=["c", "a", "b"])
+        c2 = Categorical(["b", "c"], categories=["c", "a", "b"])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical(["a", "b", "b", "c"], categories=["a", "b", "c"])
+        tm.assert_categorical_equal(result, expected)
+
+        # fastpath - skip resort
+        c1 = Categorical(["a", "b"], categories=["a", "b", "c"])
+        c2 = Categorical(["b", "c"], categories=["a", "b", "c"])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical(["a", "b", "b", "c"], categories=["a", "b", "c"])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical(["x", np.nan])
+        c2 = Categorical([np.nan, "b"])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical(["x", np.nan, np.nan, "b"], categories=["b", "x"])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical([np.nan])
+        c2 = Categorical([np.nan])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical([np.nan, np.nan])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical([])
+        c2 = Categorical([])
+        result = union_categoricals([c1, c2], sort_categories=True)
+        expected = Categorical([])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical(["b", "a"], categories=["b", "a", "c"], ordered=True)
+        c2 = Categorical(["a", "c"], categories=["b", "a", "c"], ordered=True)
+        with pytest.raises(TypeError):
+            union_categoricals([c1, c2], sort_categories=True)
+
+    def test_union_categoricals_sort_false(self):
+        # GH 13846
+        c1 = Categorical(["x", "y", "z"])
+        c2 = Categorical(["a", "b", "c"])
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical(
+            ["x", "y", "z", "a", "b", "c"], categories=["x", "y", "z", "a", "b", "c"]
+        )
+        tm.assert_categorical_equal(result, expected)
+
+        # fastpath
+        c1 = Categorical(["a", "b"], categories=["b", "a", "c"])
+        c2 = Categorical(["b", "c"], categories=["b", "a", "c"])
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical(["a", "b", "b", "c"], categories=["b", "a", "c"])
+        tm.assert_categorical_equal(result, expected)
+
+        # fastpath - skip resort
+        c1 = Categorical(["a", "b"], categories=["a", "b", "c"])
+        c2 = Categorical(["b", "c"], categories=["a", "b", "c"])
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical(["a", "b", "b", "c"], categories=["a", "b", "c"])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical(["x", np.nan])
+        c2 = Categorical([np.nan, "b"])
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical(["x", np.nan, np.nan, "b"], categories=["x", "b"])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical([np.nan])
+        c2 = Categorical([np.nan])
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical([np.nan, np.nan])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical([])
+        c2 = Categorical([])
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical([])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Categorical(["b", "a"], categories=["b", "a", "c"], ordered=True)
+        c2 = Categorical(["a", "c"], categories=["b", "a", "c"], ordered=True)
+        result = union_categoricals([c1, c2], sort_categories=False)
+        expected = Categorical(
+            ["b", "a", "a", "c"], categories=["b", "a", "c"], ordered=True
+        )
+        tm.assert_categorical_equal(result, expected)
+
+    def test_union_categorical_unwrap(self):
+        # GH 14173
+        c1 = Categorical(["a", "b"])
+        c2 = pd.Series(["b", "c"], dtype="category")
+        result = union_categoricals([c1, c2])
+        expected = Categorical(["a", "b", "b", "c"])
+        tm.assert_categorical_equal(result, expected)
+
+        c2 = CategoricalIndex(c2)
+        result = union_categoricals([c1, c2])
+        tm.assert_categorical_equal(result, expected)
+
+        c1 = Series(c1)
+        result = union_categoricals([c1, c2])
+        tm.assert_categorical_equal(result, expected)
+
+        with pytest.raises(TypeError):
+            union_categoricals([c1, ["a", "b", "c"]])

venv/lib/python3.6/site-packages/pandas/tests/reshape/test_util.py

@@ -0,0 +1,51 @@
+import numpy as np
+import pytest
+
+from pandas import Index, date_range
+from pandas.core.reshape.util import cartesian_product
+import pandas.util.testing as tm
+
+
+class TestCartesianProduct:
+    def test_simple(self):
+        x, y = list("ABC"), [1, 22]
+        result1, result2 = cartesian_product([x, y])
+        expected1 = np.array(["A", "A", "B", "B", "C", "C"])
+        expected2 = np.array([1, 22, 1, 22, 1, 22])
+        tm.assert_numpy_array_equal(result1, expected1)
+        tm.assert_numpy_array_equal(result2, expected2)
+
+    def test_datetimeindex(self):
+        # regression test for GitHub issue #6439
+        # make sure that the ordering on datetimeindex is consistent
+        x = date_range("2000-01-01", periods=2)
+        result1, result2 = [Index(y).day for y in cartesian_product([x, x])]
+        expected1 = Index([1, 1, 2, 2])
+        expected2 = Index([1, 2, 1, 2])
+        tm.assert_index_equal(result1, expected1)
+        tm.assert_index_equal(result2, expected2)
+
+    def test_empty(self):
+        # product of empty factors
+        X = [[], [0, 1], []]
+        Y = [[], [], ["a", "b", "c"]]
+        for x, y in zip(X, Y):
+            expected1 = np.array([], dtype=np.asarray(x).dtype)
+            expected2 = np.array([], dtype=np.asarray(y).dtype)
+            result1, result2 = cartesian_product([x, y])
+            tm.assert_numpy_array_equal(result1, expected1)
+            tm.assert_numpy_array_equal(result2, expected2)
+
+        # empty product (empty input):
+        result = cartesian_product([])
+        expected = []
+        assert result == expected
+
+    @pytest.mark.parametrize(
+        "X", [1, [1], [1, 2], [[1], 2], "a", ["a"], ["a", "b"], [["a"], "b"]]
+    )
+    def test_invalid_input(self, X):
+        msg = "Input must be a list-like of list-likes"
+
+        with pytest.raises(TypeError, match=msg):
+            cartesian_product(X=X)