ENH: Support scalar condition in `Series.where` and `DataFrame.where`

[randolf-scholz]

Feature Type

• Changing existing functionality in pandas

Problem Description

For compatibility reasons with numpy.where it would be nice to allow scalar booleans as a condition to Series.where and DataFrame.where. Currently

import pandas as pd
s = pd.Series([1,2])
s.where(True)

raises ValueError: Array conditional must be same shape as self

Feature Description

Allow scalar boolean as input to

• Series.where
• Series.mask
• DataFrame.where
• DataFrame.mask

Alternative Solutions

idk.

Additional Context

Regarding documentation, I am not sure anything needs to be updated since both Series.where and DataFrame.where specify array-like as a valid input type, and scalars are considered array-like.

[lithomas1]

Sure, seems reasonable to me.

[randolf-scholz]

@lithomas1 The PR has passed all tests. Currently I added to the changelog of 2.1.0, should I use 2.0.3 instead?

[lithomas1]

No, I don't think this should be backported, 2.1.0 seems fine to me (its gonna come out around in a month or so).