BUG FIX: Using Series.str.fullmatch() and Series.str.match() with a compiled regex fails with arrow strings #61964
Conversation
There was a problem hiding this comment.
Thanks!
It seems we don't actually document that we support a compiled regular expression, although it works in practice because we pass pat to re.compile() in the non-arrow version, and that works.
But so it would be good to update the documentation and typing then to reflect the fact that a compiled pattern is also supported.
|
@jorisvandenbossche Moved tests to |
The suggestions of @yuanx749 are in the good direction |
| if isinstance(pat, re.Pattern): | ||
| # GH#61952 | ||
| pat = pat.pattern | ||
| if not pat.startswith("^"): |
There was a problem hiding this comment.
This causes some typing issues (see https://github.com/pandas-dev/pandas/actions/runs/16615325282/job/47006632529?pr=61964).
The type checking does not understand that pat is now a string after doing pat.pattern ...
I suppose adding something like cast(str, pat) should fix it
See https://pandas.pydata.org/docs/dev/development/contributing_codebase.html#validating-type-hints
There was a problem hiding this comment.
@jorisvandenbossche is it ok if I add str check in the if condition, I think this will also fix the issue?:
if isinstance(pat, str) and not pat.startswith("^"):
Fixes: #61952
After Fix:
Series.str.fullmatch()andSeries.str.match()with a compiled regex fails with arrow strings #61952