TST: Add tests for single level indexing with loc(axis=1) #29519
Conversation
ryankarlos force-pushed the TST-loc-axis1-singlelevel branch from 8c0c56d to e8a27bb Compare 
| df.loc(axis="foo")[:, :, ["C1", "C3"]] | ||
| | ||
| def test_loc_axis_single_level_indexer(self): | ||
| |
There was a problem hiding this comment.
Could you add the issue number?
| result = df1.loc(axis=1)["a1"] | ||
| expected = df1.iloc[:, :3] | ||
| expected.columns = ["b1", "b2", "b3"] | ||
| tm.assert_frame_equal(result, expected) |
There was a problem hiding this comment.
Could you make 3 different functions for these tests?
There was a problem hiding this comment.
@mroeschke Is it ok if I split it into 2 functions - one for Multindex column df and one for simple df --- and for the Multiindex bit I can comment on the two different tests - i.e one is slicing on multi columns on single level and other is slicing on single col. Otherwise, having concise names for the functions is tricky.
Something like this:
# GH29519 # test with single level multiple columns slice df1 = pd.DataFrame( np.arange(27).reshape(3, 9), columns=pd.MultiIndex.from_product( [["a1", "a2", "a3"], ["b1", "b2", "b3"]] ), ) result = df1.loc(axis=1)["a1":"a2"] expected = df1.iloc[:, :-3] tm.assert_frame_equal(result, expected) # GH29519 # test with single level single column slice result = df1.loc(axis=1)["a1"] expected = df1.iloc[:, :3] expected.columns = ["b1", "b2", "b3"] tm.assert_frame_equal(result, expected) There was a problem hiding this comment.
Verbose function names are fine. Helps isolate the failing test as separate methods.
There was a problem hiding this comment.
Ok just pushed with changes as you originally suggested
| df2 = pd.DataFrame(np.arange(9).reshape(3, 3), columns=["a", "b", "c"]) | ||
| result = df2.loc(axis=1)["a"] | ||
| expected = pd.Series([0, 3, 6], name="a") | ||
| tm.assert_series_equal(result, expected, check_dtype=False) |
There was a problem hiding this comment.
Why is check_dtype=False set?
There was a problem hiding this comment.
@mroeschke Since CI was failing for some of the 32 bit OS builds because of int32 vs int64 dtype comparison - im having a similar issue in #29522. Don't know if there is any other work around.
================================== FAILURES =================================== __________ TestMultiIndexSlicers.test_loc_axis_single_level_indexer ___________ [gw0] win32 -- Python 3.6.7 C:\Miniconda\envs\pandas-dev\python.exe self = <pandas.tests.indexing.multiindex.test_slice.TestMultiIndexSlicers object at 0x00000156C517C898> def test_loc_axis_single_level_indexer(self): # test single level indexing on Multindex column dataframe df1 = pd.DataFrame( np.arange(27).reshape(3, 9), columns=pd.MultiIndex.from_product( [["a1", "a2", "a3"], ["b1", "b2", "b3"]] ), ) result = df1.loc(axis=1)["a1":"a2"] expected = df1.iloc[:, :-3] tm.assert_frame_equal(result, expected) result = df1.loc(axis=1)["a1"] expected = df1.iloc[:, :3] expected.columns = ["b1", "b2", "b3"] tm.assert_frame_equal(result, expected) # test single level indexing on single index column data frame df2 = pd.DataFrame(np.arange(9).reshape(3, 3), columns=["a", "b", "c"]) result = df2.loc(axis=1)["a"] expected = pd.Series([0, 3, 6], dtype="int64", name="a") > tm.assert_series_equal(result, expected) E AssertionError: Attributes of Series are different E E Attribute "dtype" are different E [left]: int32 E [right]: int64 pandas\tests\indexing\multiindex\test_slice.py:516: AssertionError There was a problem hiding this comment.
I see you're not assigning dtype="int64" anymore. Could you remove check_dtype=False now?
gfyoung added DataFrame 
| @mroeschke all green now |
| thanks @ryankarlos |
4 participants
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diff