ENH: Allow export of mixed columns to Stata strl

[bashtage]

Enable export of large columns to Stata strls when the column
contains None as a null value

closes #23633

• closes StataWriter for version 117 fails on None in a string column long enough to be a Stata StrL. #23633
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[pep8speaks]

Hello @bashtage! Thanks for submitting the PR.

• There are no PEP8 issues in the file pandas/io/stata.py !

• There are no PEP8 issues in the file pandas/tests/io/test_stata.py !

[codecov]

Codecov Report

Merging #23692 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master   #23692   +/-   ##
=======================================
  Coverage   92.24%   92.24%           
=======================================
  Files         161      161           
  Lines       51318    51318           
=======================================
  Hits        47339    47339           
  Misses       3979     3979

Flag	Coverage Δ
#multiple	90.63% <ø> (ø)	⬆️
#single	42.31% <ø> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a197837...75f9d80. Read the comment docs.

[jreback]

does this break idempotency? (i think answer is no, this is round-trippable)

[kylebarron]

I think yes because Stata doesn't have a string missing value. When the Stata file is read back in to Pandas, the None values are ''.

This already happens with shorter strings with the Stata 114 writer. This PR allows the same to happen with strings longer than 245 characters in the 117 writer.

>>> import pandas as pd
>>> df = pd.DataFrame({'a': ['abc', None]})
>>> df.to_stata('test.dta')
>>> pd.read_stata('test.dta')

Out[7]:
   index    a
0      0  abc
1      1

[jreback]

is it worth having a test for other versions?

[bashtage]

Not relevant for the other version (114) which doesn't support strls.

[bashtage]

does this break idempotency? (i think the answer is no, this is round-trippable)

Yes. Essentially you get None->'' conversion in mixed columns since there is no missing value for strings in Stata.

The only way to have idempotency would be to raise on mixed string columns so that None is not allowed in string columns. Then users would need to convert None to '' before saving, so that '' -> file -> ''.

[jreback]

thanks @bashtage