TST: added test for pd.where overflow error GH31687 #49926
Conversation
|
|
||
|
|
||
| @pytest.mark.parametrize( | ||
| "replacement", [0.001, True, "snake", DATETIME_JAN_1_1900_OPTIONAL_TZ] |
There was a problem hiding this comment.
DATETIME_JAN_1_1900_OPTIONAL_TZ is gonna be a LazyStrategy object here, which I guess isn't quite what you wanted, I think you'll need @given to use hypothesis tests
Not sure we need that though, just some date should be fine if you want to test using a date as a replacement
Also, the original issue used None, shall we include that?
There was a problem hiding this comment.
You're right, I didn't want a LazyStrategy object! I'll change that to a date, and also include None to align better with the original issue. I was also wondering if there was already a quick way to test all scalar types rather than having to use a long parametrized list. Or would that be overkill?
There was a problem hiding this comment.
yeah I think just keeping the original snippet from the issue as a test should be fine
| ) | ||
| def test_where_int_overflow(replacement): | ||
| # GH 31687 | ||
| df = DataFrame([[1.0, 2e19, "nine"], [np.nan, 0.1, None]]) |
There was a problem hiding this comment.
ideally for regression tests it's better to keep them as close as possible to the original - any reason to use 2e19 instead of 2e25?
There was a problem hiding this comment.
I think I reduced it down to the smallest order of magnitude that still had the problem in pandas 1.4.1, but I don't have strong feelings about using that instead of the 2e25 from the original.
abc954c to
48f81cd
Compare
Added a test for pd.where to /indexing/test_where.py
doc/source/whatsnew/vX.X.X.rstfile if fixing a bug or adding a new feature.