Incorrect suggestion to replace == True with is True for pandas dataframe

[unknown-eps]

Summary

https://play.ruff.rs/c4303e07-49f1-456d-b48c-6a6a9b9bfcd5

import pandas as pd

temp =pd.DataFrame({'a':[1,2,3],'b':['True','True','False']})

filter_temp = temp[temp['b'] == True] #Correct

ruff_suggested = temp[temp['b'] is True] # Will give error on running

Ruff suggests the second one

Version

No response

[Daverball]

This requires type inference and multi-file analysis, both of which are a little ways off still.

This boils down to, that the rule should not trigger when x == True returns something other than bool.

I tend to use .eq(True) instead of == True in pandas, it reads better and won't trigger this rule. Usually you don't even really need that though and can just directly use temp['b'] as the filter, since eq(True) just gives you back the same series, if all the values are already boolean, so you only need it when you have mixed values.

[dylwil3]

Sorry you ran into this! This is a known issue, and it's why the rule auto-fix is marked as unsafe (see the documentation for the rule)