def test_join_indexed_dataframe_to_indexed_dataframe():
A = pd.DataFrame({'x': [1, 2, 3, 4, 5, 6]}, index=[1, 2, 3, 4, 6, 7])
a = dd.repartition(A, [1, 4, 7])
B = pd.DataFrame({'y': list('abcdef')}, index=[1, 2, 4, 5, 6, 8])
b = dd.repartition(B, [1, 2, 5, 8])
c = join_indexed_dataframes(a, b, how='left')
assert c.divisions[0] == a.divisions[0]
assert c.divisions[-1] == a.divisions[-1]
tm.assert_frame_equal(c.compute(), A.join(B))
c = join_indexed_dataframes(a, b, how='right')
assert c.divisions[0] == b.divisions[0]
assert c.divisions[-1] == b.divisions[-1]
tm.assert_frame_equal(c.compute(), A.join(B, how='right'))
c = join_indexed_dataframes(a, b, how='inner')
assert c.divisions[0] == 1
assert c.divisions[-1] == 7
tm.assert_frame_equal(c.compute(), A.join(B, how='inner'))
c = join_indexed_dataframes(a, b, how='outer')
assert c.divisions[0] == 1
assert c.divisions[-1] == 8
tm.assert_frame_equal(c.compute(), A.join(B, how='outer'))
def test_join_indexed_dataframe_to_indexed_dataframe():
A = pd.DataFrame({'x': [1, 2, 3, 4, 5, 6]},
index=[1, 2, 3, 4, 6, 7])
a = dd.repartition(A, [1, 4, 7])
B = pd.DataFrame({'y': list('abcdef')},
index=[1, 2, 4, 5, 6, 8])
b = dd.repartition(B, [1, 2, 5, 8])
c = join_indexed_dataframes(a, b, how='left')
assert c.divisions[0] == a.divisions[0]
assert c.divisions[-1] == max(a.divisions + b.divisions)
assert eq(c, A.join(B))
c = join_indexed_dataframes(a, b, how='right')
assert c.divisions[0] == b.divisions[0]
assert c.divisions[-1] == b.divisions[-1]
assert eq(c, A.join(B, how='right'))
c = join_indexed_dataframes(a, b, how='inner')
assert c.divisions[0] == 1
assert c.divisions[-1] == max(a.divisions + b.divisions)
assert eq(c.compute(), A.join(B, how='inner'))
c = join_indexed_dataframes(a, b, how='outer')
assert c.divisions[0] == 1
assert c.divisions[-1] == 8
assert eq(c.compute(), A.join(B, how='outer'))
assert sorted(join_indexed_dataframes(a, b, how='inner').dask) == \
sorted(join_indexed_dataframes(a, b, how='inner').dask)
assert sorted(join_indexed_dataframes(a, b, how='inner').dask) != \
sorted(join_indexed_dataframes(a, b, how='outer').dask)
def test_join_indexed_dataframe_to_indexed_dataframe():
A = pd.DataFrame({"x": [1, 2, 3, 4, 5, 6]}, index=[1, 2, 3, 4, 6, 7])
a = dd.repartition(A, [1, 4, 7])
B = pd.DataFrame({"y": list("abcdef")}, index=[1, 2, 4, 5, 6, 8])
b = dd.repartition(B, [1, 2, 5, 8])
c = join_indexed_dataframes(a, b, how="left")
assert c.divisions[0] == a.divisions[0]
assert c.divisions[-1] == max(a.divisions + b.divisions)
assert eq(c, A.join(B))
c = join_indexed_dataframes(a, b, how="right")
assert c.divisions[0] == b.divisions[0]
assert c.divisions[-1] == b.divisions[-1]
assert eq(c, A.join(B, how="right"))
c = join_indexed_dataframes(a, b, how="inner")
assert c.divisions[0] == 1
assert c.divisions[-1] == max(a.divisions + b.divisions)
assert eq(c.compute(), A.join(B, how="inner"))
c = join_indexed_dataframes(a, b, how="outer")
assert c.divisions[0] == 1
assert c.divisions[-1] == 8
assert eq(c.compute(), A.join(B, how="outer"))
assert sorted(join_indexed_dataframes(a, b, how="inner").dask) == sorted(
join_indexed_dataframes(a, b, how="inner").dask
)
assert sorted(join_indexed_dataframes(a, b, how="inner").dask) != sorted(
join_indexed_dataframes(a, b, how="outer").dask
)