Docs: Describe csv.reader conversions for all quoting modes

[esadomer]

Summary

Update the csv.reader overview to describe conversions performed by QUOTE_STRINGS and QUOTE_NOTNULL. It currently says all fields are strings unless QUOTE_NONNUMERIC is used, contradicting the descriptions of the newer quoting modes.

import csv

row = ',3,"5","",'
assert next(csv.reader([row], quoting=csv.QUOTE_STRINGS)) == [None, 3.0, "5", "", None]
assert next(csv.reader([row], quoting=csv.QUOTE_NOTNULL)) == [None, "3", "5", "", None]

The revised text distinguishes non-empty unquoted fields converted to floats from empty unquoted fields converted to None.

Validation

• Built CPython 3.16.0a0 from the checkout with PCbuild\build.bat -p x64 -d -M on Windows 10 x64 (MSVC 19.30).
• PCbuild\amd64\python_d.exe -m test test_csv: SUCCESS (run=156, skipped=0).
• Verified the examples and boundary cases with the same debug interpreter.
• python -m sphinx -W --keep-going -j 1 -b html -d build/doctrees . build/html from Doc: incremental HTML build passed with Sphinx 8.2.3. The clean baseline build reported an existing duplicate customization label in using/windows.rst and reference/datamodel.rst.
• python -m sphinxlint Doc/library/csv.rst, PCbuild\amd64\python_d.exe Tools\patchcheck\patchcheck.py, and git diff --check: passed.

This changes documentation only. No NEWS entry is needed. The full interpreter test suite was not run.

[read-the-docs-community]

Documentation build overview

📚 cpython-previews | 🛠️ Build #34831752 | 📁 Comparing 13b288e against main (84d4601)

  🔍 Preview build  

1 file changed

± library/csv.html