306 lines
11 KiB
Python
306 lines
11 KiB
Python
|
import numpy as np
|
||
|
|
import pytest
|
||
|
|
import pytz
|
||
|
|
|
||
|
|
import pandas as pd
|
||
|
|
from pandas import DataFrame, Index, Series, date_range
|
||
|
|
import pandas._testing as tm
|
||
|
|
|
||
|
|
|
||
|
|
class TestDataFrameAlign:
|
||
|
|
def test_frame_align_aware(self):
|
||
|
|
idx1 = date_range("2001", periods=5, freq="H", tz="US/Eastern")
|
||
|
|
idx2 = date_range("2001", periods=5, freq="2H", tz="US/Eastern")
|
||
|
|
df1 = DataFrame(np.random.randn(len(idx1), 3), idx1)
|
||
|
|
df2 = DataFrame(np.random.randn(len(idx2), 3), idx2)
|
||
|
|
new1, new2 = df1.align(df2)
|
||
|
|
assert df1.index.tz == new1.index.tz
|
||
|
|
assert df2.index.tz == new2.index.tz
|
||
|
|
|
||
|
|
# different timezones convert to UTC
|
||
|
|
|
||
|
|
# frame with frame
|
||
|
|
df1_central = df1.tz_convert("US/Central")
|
||
|
|
new1, new2 = df1.align(df1_central)
|
||
|
|
assert new1.index.tz == pytz.UTC
|
||
|
|
assert new2.index.tz == pytz.UTC
|
||
|
|
|
||
|
|
# frame with Series
|
||
|
|
new1, new2 = df1.align(df1_central[0], axis=0)
|
||
|
|
assert new1.index.tz == pytz.UTC
|
||
|
|
assert new2.index.tz == pytz.UTC
|
||
|
|
|
||
|
|
df1[0].align(df1_central, axis=0)
|
||
|
|
assert new1.index.tz == pytz.UTC
|
||
|
|
assert new2.index.tz == pytz.UTC
|
||
|
|
|
||
|
|
def test_align_float(self, float_frame):
|
||
|
|
af, bf = float_frame.align(float_frame)
|
||
|
|
assert af._mgr is not float_frame._mgr
|
||
|
|
|
||
|
|
af, bf = float_frame.align(float_frame, copy=False)
|
||
|
|
assert af._mgr is float_frame._mgr
|
||
|
|
|
||
|
|
# axis = 0
|
||
|
|
other = float_frame.iloc[:-5, :3]
|
||
|
|
af, bf = float_frame.align(other, axis=0, fill_value=-1)
|
||
|
|
|
||
|
|
tm.assert_index_equal(bf.columns, other.columns)
|
||
|
|
|
||
|
|
# test fill value
|
||
|
|
join_idx = float_frame.index.join(other.index)
|
||
|
|
diff_a = float_frame.index.difference(join_idx)
|
||
|
|
diff_b = other.index.difference(join_idx)
|
||
|
|
diff_a_vals = af.reindex(diff_a).values
|
||
|
|
diff_b_vals = bf.reindex(diff_b).values
|
||
|
|
assert (diff_a_vals == -1).all()
|
||
|
|
|
||
|
|
af, bf = float_frame.align(other, join="right", axis=0)
|
||
|
|
tm.assert_index_equal(bf.columns, other.columns)
|
||
|
|
tm.assert_index_equal(bf.index, other.index)
|
||
|
|
tm.assert_index_equal(af.index, other.index)
|
||
|
|
|
||
|
|
# axis = 1
|
||
|
|
other = float_frame.iloc[:-5, :3].copy()
|
||
|
|
af, bf = float_frame.align(other, axis=1)
|
||
|
|
tm.assert_index_equal(bf.columns, float_frame.columns)
|
||
|
|
tm.assert_index_equal(bf.index, other.index)
|
||
|
|
|
||
|
|
# test fill value
|
||
|
|
join_idx = float_frame.index.join(other.index)
|
||
|
|
diff_a = float_frame.index.difference(join_idx)
|
||
|
|
diff_b = other.index.difference(join_idx)
|
||
|
|
diff_a_vals = af.reindex(diff_a).values
|
||
|
|
|
||
|
|
# TODO(wesm): unused?
|
||
|
|
diff_b_vals = bf.reindex(diff_b).values # noqa
|
||
|
|
|
||
|
|
assert (diff_a_vals == -1).all()
|
||
|
|
|
||
|
|
af, bf = float_frame.align(other, join="inner", axis=1)
|
||
|
|
tm.assert_index_equal(bf.columns, other.columns)
|
||
|
|
|
||
|
|
af, bf = float_frame.align(other, join="inner", axis=1, method="pad")
|
||
|
|
tm.assert_index_equal(bf.columns, other.columns)
|
||
|
|
|
||
|
|
af, bf = float_frame.align(
|
||
|
|
other.iloc[:, 0], join="inner", axis=1, method=None, fill_value=None
|
||
|
|
)
|
||
|
|
tm.assert_index_equal(bf.index, Index([]))
|
||
|
|
|
||
|
|
af, bf = float_frame.align(
|
||
|
|
other.iloc[:, 0], join="inner", axis=1, method=None, fill_value=0
|
||
|
|
)
|
||
|
|
tm.assert_index_equal(bf.index, Index([]))
|
||
|
|
|
||
|
|
# Try to align DataFrame to Series along bad axis
|
||
|
|
msg = "No axis named 2 for object type DataFrame"
|
||
|
|
with pytest.raises(ValueError, match=msg):
|
||
|
|
float_frame.align(af.iloc[0, :3], join="inner", axis=2)
|
||
|
|
|
||
|
|
# align dataframe to series with broadcast or not
|
||
|
|
idx = float_frame.index
|
||
|
|
s = Series(range(len(idx)), index=idx)
|
||
|
|
|
||
|
|
left, right = float_frame.align(s, axis=0)
|
||
|
|
tm.assert_index_equal(left.index, float_frame.index)
|
||
|
|
tm.assert_index_equal(right.index, float_frame.index)
|
||
|
|
assert isinstance(right, Series)
|
||
|
|
|
||
|
|
left, right = float_frame.align(s, broadcast_axis=1)
|
||
|
|
tm.assert_index_equal(left.index, float_frame.index)
|
||
|
|
expected = {c: s for c in float_frame.columns}
|
||
|
|
expected = DataFrame(
|
||
|
|
expected, index=float_frame.index, columns=float_frame.columns
|
||
|
|
)
|
||
|
|
tm.assert_frame_equal(right, expected)
|
||
|
|
|
||
|
|
# see gh-9558
|
||
|
|
df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
|
||
|
|
result = df[df["a"] == 2]
|
||
|
|
expected = DataFrame([[2, 5]], index=[1], columns=["a", "b"])
|
||
|
|
tm.assert_frame_equal(result, expected)
|
||
|
|
|
||
|
|
result = df.where(df["a"] == 2, 0)
|
||
|
|
expected = DataFrame({"a": [0, 2, 0], "b": [0, 5, 0]})
|
||
|
|
tm.assert_frame_equal(result, expected)
|
||
|
|
|
||
|
|
def test_align_int(self, int_frame):
|
||
|
|
# test other non-float types
|
||
|
|
other = DataFrame(index=range(5), columns=["A", "B", "C"])
|
||
|
|
|
||
|
|
af, bf = int_frame.align(other, join="inner", axis=1, method="pad")
|
||
|
|
tm.assert_index_equal(bf.columns, other.columns)
|
||
|
|
|
||
|
|
def test_align_mixed_type(self, float_string_frame):
|
||
|
|
|
||
|
|
af, bf = float_string_frame.align(
|
||
|
|
float_string_frame, join="inner", axis=1, method="pad"
|
||
|
|
)
|
||
|
|
tm.assert_index_equal(bf.columns, float_string_frame.columns)
|
||
|
|
|
||
|
|
def test_align_mixed_float(self, mixed_float_frame):
|
||
|
|
# mixed floats/ints
|
||
|
|
other = DataFrame(index=range(5), columns=["A", "B", "C"])
|
||
|
|
|
||
|
|
af, bf = mixed_float_frame.align(
|
||
|
|
other.iloc[:, 0], join="inner", axis=1, method=None, fill_value=0
|
||
|
|
)
|
||
|
|
tm.assert_index_equal(bf.index, Index([]))
|
||
|
|
|
||
|
|
def test_align_mixed_int(self, mixed_int_frame):
|
||
|
|
other = DataFrame(index=range(5), columns=["A", "B", "C"])
|
||
|
|
|
||
|
|
af, bf = mixed_int_frame.align(
|
||
|
|
other.iloc[:, 0], join="inner", axis=1, method=None, fill_value=0
|
||
|
|
)
|
||
|
|
tm.assert_index_equal(bf.index, Index([]))
|
||
|
|
|
||
|
|
@pytest.mark.parametrize(
|
||
|
|
"l_ordered,r_ordered,expected",
|
||
|
|
[
|
||
|
|
[True, True, pd.CategoricalIndex],
|
||
|
|
[True, False, Index],
|
||
|
|
[False, True, Index],
|
||
|
|
[False, False, pd.CategoricalIndex],
|
||
|
|
],
|
||
|
|
)
|
||
|
|
def test_align_categorical(self, l_ordered, r_ordered, expected):
|
||
|
|
# GH-28397
|
||
|
|
df_1 = DataFrame(
|
||
|
|
{
|
||
|
|
"A": np.arange(6, dtype="int64"),
|
||
|
|
"B": Series(list("aabbca")).astype(
|
||
|
|
pd.CategoricalDtype(list("cab"), ordered=l_ordered)
|
||
|
|
),
|
||
|
|
}
|
||
|
|
).set_index("B")
|
||
|
|
df_2 = DataFrame(
|
||
|
|
{
|
||
|
|
"A": np.arange(5, dtype="int64"),
|
||
|
|
"B": Series(list("babca")).astype(
|
||
|
|
pd.CategoricalDtype(list("cab"), ordered=r_ordered)
|
||
|
|
),
|
||
|
|
}
|
||
|
|
).set_index("B")
|
||
|
|
|
||
|
|
aligned_1, aligned_2 = df_1.align(df_2)
|
||
|
|
assert isinstance(aligned_1.index, expected)
|
||
|
|
assert isinstance(aligned_2.index, expected)
|
||
|
|
tm.assert_index_equal(aligned_1.index, aligned_2.index)
|
||
|
|
|
||
|
|
def test_align_multiindex(self):
|
||
|
|
# GH#10665
|
||
|
|
# same test cases as test_align_multiindex in test_series.py
|
||
|
|
|
||
|
|
midx = pd.MultiIndex.from_product(
|
||
|
|
[range(2), range(3), range(2)], names=("a", "b", "c")
|
||
|
|
)
|
||
|
|
idx = Index(range(2), name="b")
|
||
|
|
df1 = DataFrame(np.arange(12, dtype="int64"), index=midx)
|
||
|
|
df2 = DataFrame(np.arange(2, dtype="int64"), index=idx)
|
||
|
|
|
||
|
|
# these must be the same results (but flipped)
|
||
|
|
res1l, res1r = df1.align(df2, join="left")
|
||
|
|
res2l, res2r = df2.align(df1, join="right")
|
||
|
|
|
||
|
|
expl = df1
|
||
|
|
tm.assert_frame_equal(expl, res1l)
|
||
|
|
tm.assert_frame_equal(expl, res2r)
|
||
|
|
expr = DataFrame([0, 0, 1, 1, np.nan, np.nan] * 2, index=midx)
|
||
|
|
tm.assert_frame_equal(expr, res1r)
|
||
|
|
tm.assert_frame_equal(expr, res2l)
|
||
|
|
|
||
|
|
res1l, res1r = df1.align(df2, join="right")
|
||
|
|
res2l, res2r = df2.align(df1, join="left")
|
||
|
|
|
||
|
|
exp_idx = pd.MultiIndex.from_product(
|
||
|
|
[range(2), range(2), range(2)], names=("a", "b", "c")
|
||
|
|
)
|
||
|
|
expl = DataFrame([0, 1, 2, 3, 6, 7, 8, 9], index=exp_idx)
|
||
|
|
tm.assert_frame_equal(expl, res1l)
|
||
|
|
tm.assert_frame_equal(expl, res2r)
|
||
|
|
expr = DataFrame([0, 0, 1, 1] * 2, index=exp_idx)
|
||
|
|
tm.assert_frame_equal(expr, res1r)
|
||
|
|
tm.assert_frame_equal(expr, res2l)
|
||
|
|
|
||
|
|
def test_align_series_combinations(self):
|
||
|
|
df = DataFrame({"a": [1, 3, 5], "b": [1, 3, 5]}, index=list("ACE"))
|
||
|
|
s = Series([1, 2, 4], index=list("ABD"), name="x")
|
||
|
|
|
||
|
|
# frame + series
|
||
|
|
res1, res2 = df.align(s, axis=0)
|
||
|
|
exp1 = DataFrame(
|
||
|
|
{"a": [1, np.nan, 3, np.nan, 5], "b": [1, np.nan, 3, np.nan, 5]},
|
||
|
|
index=list("ABCDE"),
|
||
|
|
)
|
||
|
|
exp2 = Series([1, 2, np.nan, 4, np.nan], index=list("ABCDE"), name="x")
|
||
|
|
|
||
|
|
tm.assert_frame_equal(res1, exp1)
|
||
|
|
tm.assert_series_equal(res2, exp2)
|
||
|
|
|
||
|
|
# series + frame
|
||
|
|
res1, res2 = s.align(df)
|
||
|
|
tm.assert_series_equal(res1, exp2)
|
||
|
|
tm.assert_frame_equal(res2, exp1)
|
||
|
|
|
||
|
|
def _check_align(self, a, b, axis, fill_axis, how, method, limit=None):
|
||
|
|
aa, ab = a.align(
|
||
|
|
b, axis=axis, join=how, method=method, limit=limit, fill_axis=fill_axis
|
||
|
|
)
|
||
|
|
|
||
|
|
join_index, join_columns = None, None
|
||
|
|
|
||
|
|
ea, eb = a, b
|
||
|
|
if axis is None or axis == 0:
|
||
|
|
join_index = a.index.join(b.index, how=how)
|
||
|
|
ea = ea.reindex(index=join_index)
|
||
|
|
eb = eb.reindex(index=join_index)
|
||
|
|
|
||
|
|
if axis is None or axis == 1:
|
||
|
|
join_columns = a.columns.join(b.columns, how=how)
|
||
|
|
ea = ea.reindex(columns=join_columns)
|
||
|
|
eb = eb.reindex(columns=join_columns)
|
||
|
|
|
||
|
|
ea = ea.fillna(axis=fill_axis, method=method, limit=limit)
|
||
|
|
eb = eb.fillna(axis=fill_axis, method=method, limit=limit)
|
||
|
|
|
||
|
|
tm.assert_frame_equal(aa, ea)
|
||
|
|
tm.assert_frame_equal(ab, eb)
|
||
|
|
|
||
|
|
@pytest.mark.parametrize("meth", ["pad", "bfill"])
|
||
|
|
@pytest.mark.parametrize("ax", [0, 1, None])
|
||
|
|
@pytest.mark.parametrize("fax", [0, 1])
|
||
|
|
@pytest.mark.parametrize("how", ["inner", "outer", "left", "right"])
|
||
|
|
def test_align_fill_method(self, how, meth, ax, fax, float_frame):
|
||
|
|
df = float_frame
|
||
|
|
self._check_align_fill(df, how, meth, ax, fax)
|
||
|
|
|
||
|
|
def _check_align_fill(self, frame, kind, meth, ax, fax):
|
||
|
|
left = frame.iloc[0:4, :10]
|
||
|
|
right = frame.iloc[2:, 6:]
|
||
|
|
empty = frame.iloc[:0, :0]
|
||
|
|
|
||
|
|
self._check_align(left, right, axis=ax, fill_axis=fax, how=kind, method=meth)
|
||
|
|
self._check_align(
|
||
|
|
left, right, axis=ax, fill_axis=fax, how=kind, method=meth, limit=1
|
||
|
|
)
|
||
|
|
|
||
|
|
# empty left
|
||
|
|
self._check_align(empty, right, axis=ax, fill_axis=fax, how=kind, method=meth)
|
||
|
|
self._check_align(
|
||
|
|
empty, right, axis=ax, fill_axis=fax, how=kind, method=meth, limit=1
|
||
|
|
)
|
||
|
|
|
||
|
|
# empty right
|
||
|
|
self._check_align(left, empty, axis=ax, fill_axis=fax, how=kind, method=meth)
|
||
|
|
self._check_align(
|
||
|
|
left, empty, axis=ax, fill_axis=fax, how=kind, method=meth, limit=1
|
||
|
|
)
|
||
|
|
|
||
|
|
# both empty
|
||
|
|
self._check_align(empty, empty, axis=ax, fill_axis=fax, how=kind, method=meth)
|
||
|
|
self._check_align(
|
||
|
|
empty, empty, axis=ax, fill_axis=fax, how=kind, method=meth, limit=1
|
||
|
|
)
|