TEST: Add test for select_dtypes on an empty DataFrame #62359
Conversation
deepith-18
force-pushed
the
test-select-dtypes-empty
branch
from
3e19eeb to
59712b5
Compare
| If 0 or 'index' counts are generated for each column. | ||
| If 1 or 'columns' counts are generated for each row. | ||
| numeric_only : bool, default False | ||
| numeric_only : bool, default False. |
There was a problem hiding this comment.
this needs to be reverted?
| result.iloc[0, 0] = 0 | ||
| tm.assert_frame_equal(df, df_orig) | ||
|
|
||
| def test_select_dtypes_empty_frame(): |
There was a problem hiding this comment.
please explain why this test is needed. I assume you have identified a gap in the testing, perhaps using code coverage tools or found an reported issue on the pandas issue tracker which is fixed on main and labelled Needs Tests to prevent regression?
A separate test is usually added at the end of the file when adding a regression test with a link to a known issue.
For general testing, the input dataframe could be added to the parameterisation used in the existing tests.
Or if having a dedicated test for an empty DataFrame, we would probably also want to test more than just the str type representation of a single dtype to include. For full coverage, we would want to test more dtypes, list like of dtypes and the same for the exclude parameter?
There was a problem hiding this comment.
Thanks for the detailed review and guidance, @simonjayhawkins! I will revert the change in frame.py.
My apologies for adding a separate test function. I'm new to the testing suite and your suggestion to add the empty DataFrame to the existing parameterized tests makes a lot of sense. I will work on that now.
in that case, maybe better to, in the first instance, have a go at addressing an open issue that is tagged Needs Tests and Good First Issue. I'll close this PR for now. |
2 participants
This PR adds a new test to verify the behavior of
DataFrame.select_dtypeswhen called on an empty DataFrame.The test confirms that the function correctly returns an empty DataFrame with the original index, which is the expected behavior.