DEPR: Deprecate tupleize_cols in read_csv

[gfyoung]

xref #17060 (comment)

[codecov]

Codecov Report

Merging #17820 into master will decrease coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #17820      +/-   ##
==========================================
- Coverage   91.26%   91.24%   -0.02%     
==========================================
  Files         163      163              
  Lines       49980    49982       +2     
==========================================
- Hits        45613    45606       -7     
- Misses       4367     4376       +9

Flag	Coverage Δ
#multiple	89.05% <100%> (ø)	⬆️
#single	40.25% <100%> (-0.06%)	⬇️

Impacted Files	Coverage Δ
pandas/io/parsers.py	95.5% <100%> (ø)	⬆️
pandas/io/gbq.py	25% <0%> (-58.34%)	⬇️
pandas/core/frame.py	97.74% <0%> (-0.1%)	⬇️

---

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 9f0ee53...ee4e218. Read the comment docs.

[codecov]

Codecov Report

Merging #17820 into master will decrease coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #17820      +/-   ##
==========================================
- Coverage   91.24%   91.22%   -0.02%     
==========================================
  Files         163      163              
  Lines       49992    49994       +2     
==========================================
- Hits        45613    45606       -7     
- Misses       4379     4388       +9

Flag	Coverage Δ
#multiple	89.02% <100%> (ø)	⬆️
#single	40.24% <100%> (-0.06%)	⬇️

Impacted Files	Coverage Δ
pandas/io/parsers.py	95.5% <100%> (ø)	⬆️
pandas/io/gbq.py	25% <0%> (-58.34%)	⬇️
pandas/core/frame.py	97.74% <0%> (-0.1%)	⬇️

---

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 2ebe085...bee9012. Read the comment docs.

[jreback]

you can say that this will always convert to a MI

[gfyoung]

Yep, done.

[jreback]

minor comment. merge when ready.

[jorisvandenbossche]

@gfyoung The warning does not seem to appear when doing read_csv(..., tupleize_cols=False), so I think we need to put the default to None, so we can warn in those cases as well?

[gfyoung]

@jorisvandenbossche : That's deliberate. We only issue a warning on non-default arguments. Also, tupleize_cols=False in the context read_csv means return an MI, so we're perfectly fine.

[jorisvandenbossche]

Why do we do that deliberately?
Nothing prevents a user to specify a default value, and eg to write pd.read_csv(..., tupleize_cols=False). This code will then break in the future without any warning.

[jorisvandenbossche]

We only issue a warning on default arguments.

You mean 'non-default' arguments?

[gfyoung]

You mean 'non-default' arguments?

Oops, yes 😄

Why do we do that deliberately?
Nothing prevents a user to specify a default value, and eg to write pd.read_csv(..., tupleize_cols=False).

That's possible. My concern is that's going to be confusing to a user if they see tupleize_cols=None in the docs. This deprecation methodology that I'm using is implemented throughout the codebase.

That being said, you do bring up an interesting point. While I haven't seen that problem arise with any of our other deprecations, I don't see why we shouldn't discuss it.

[gfyoung]

@jorisvandenbossche : I opened #17828 to address your point about function argument deprecations.

[jorisvandenbossche]

I disagree with not raising for the default value is how it is typically done. I think it is rather the other way around (we normally put the deprecated one at None, to be able to distinguish the default and an actual user passed value), but will answer on the issue you created.