[Python] How to group by multiple columns and apply a custom function in pandas?

I have a DataFrame in pandas with multiple columns like ['category', 'sub_category', 'value'], and I want to group by both category and sub_category, and then apply a custom function to the value column (e.g. weighted average or a conditional calculation).

I’m not sure how to properly use groupby().apply() or agg() for this. Could someone show the cleanest way to do this with an example?

I tried this:

df.groupby(['category', 'sub_category'])['value'].apply(my_custom_function)


But I got an error when my function expected more than just the Series.

Continue reading...