gh-155097: Add converter and formatter parameters in csv.reader and csv.writer

[serhiy-storchaka]

The converter parameter of csv.reader is used instead of float() for converting unquoted fields, and the formatter parameter of csv.writer is used instead of str() for converting values to strings. Both are called with the 0-based position of the field in the row as the first argument, so the conversion can depend on the column.

• Issue: Add converter and formatter parameters to csv.reader and csv.writer #155097

[read-the-docs-community]

Documentation build overview

📚 cpython-previews | 🛠️ Build #34517335 | 📁 Comparing c5f3db1 against main (204feba)

  🔍 Preview build  

128 files changed · + 1 added · ± 127 modified

+ Added

• library/time-complexity.html

± Modified

• glossary.html
• index.html
• c-api/arg.html
• c-api/capsule.html
• c-api/complex.html
• c-api/frame.html
• c-api/import.html
• c-api/marshal.html
• c-api/module.html
• c-api/type.html
• and 117 more...

[serhiy-storchaka]

Could you please look at this @smontanaro?

[smontanaro]

@serhiy-storchaka It looks fine to me. I would add the little bit of motivation you mentioned above to the descriptions of the converter and formatter arguments:

Both are called with the 0-based position of the field in the row as the first argument, so the conversion can depend on the column.

And in your original issue:

The index goes first, like in enumerate(). This also makes a wrong one-argument callable fail at once: converter=int raises TypeError on the first field instead of taking the index as the base.

I personally don't see an example where this would be useful to me. Maybe an example would be useful (I guess your money() function in the original issue would work?). Still, if this is convention for other modules' reading and writing stuff, maybe everybody already sees the value.

[serhiy-storchaka]

Thank you. Add the bit of motivation (why index is here) and examples. Rationale for the order of arguments is for reviewers, not end users. Users will just accept it.

And yes, the wrappers around the csv module can provide more sophisticated features -- currently they just cannot, the formatter and the converter were hardcoded.