BUG: replace with regex raising for StringDType

[mzeitlin11]

• closes BUG: replacement works for object but not string dtype #35977, closes BUG: Unexpected behavior pandas.DataFrame.replace with "string" dtype #41333
• tests added / passed
• Ensure all linting tests pass, see here for how to run them
• whatsnew entry

2 separate issues at play here - regex was ignored in #41333 (specific to the replace_list path) but then if replace_regex was hit with anything but an object type that would just raise.

[simonjayhawkins]

Thanks @mzeitlin11 generally lgtm for the fix in hand, although we should probably be dispatching to the _str_replace method on the EA.

[mzeitlin11]

Thanks @mzeitlin11 generally lgtm for the fix in hand, although we should probably be dispatching to the _str_replace method on the EA.

Agreed, but I think that decision could be handled independently of fixing this. There's some additional complexity because I think _str_replace assumes replacement occurs with a string, which is not guaranteed by the general replace function (though it would be a strange thing to do).

[simonjayhawkins]

Agreed, but I think that decision could be handled independently of fixing this.

sure.

[simonjayhawkins]

lgtm pending green

[jreback]

thanks @mzeitlin11 very nice!