[mraviraz]

Addresses #3009

Replace DataFrame operations using inplace=True with assignments to the returned DataFrame across eight modules. This addresses the Pandas 3.1 deprecation warnings while preserving each transformation order.

• 29 calls updated
• Tests not run

[ValueRaider]

git squash

[arhancanli]

Checked this on the branch (f5764de) against master with pandas 3.x on Python 3.12. I turned warnings into errors (warnings.simplefilter('error')) and ran utils.auto_adjust, utils.back_adjust and utils.parse_actions on a small synthetic frame and event payload: master raises the Pandas4Warning on set_index/sort_index, the branch runs clean, and the outputs are identical (same columns, same adjusted Close, same Dividends/Stock Splits frames). grep -rn inplace yfinance/ on the branch returns nothing, so no call was missed.

On whether any caller relied on the in-place mutation: auto_adjust/back_adjust work on data.copy() and the only callers in scrapers/history.py reassign (df = utils.auto_adjust(df)), so that one is safe. I only exercised the utils paths, not the scraper modules (holders, calendars, multi), which need network or fixtures; the full test suite was not run.

[apoptosis66]

Yes please, sooner than later.

[ValueRaider]

This isn't merging without a git squash. Someone take over

[apoptosis66]

squash with what? What other PRs do you want it squashed with?

[apoptosis66]

At the risk of sounding dumb. I tried to do as you suggested and squash into a new pr. I just don't know what I am doing with git/github. Sorry I am 20 year svn user. :) When I was done I had a pr that had way more commits in it than the original 8. Someone else is going to have to do it.

[ValueRaider]

Drop the AI slop. Just link to the docs https://ranaroussi.github.io/yfinance/development/code.html