filter dataframe after groupby and nunique in pand

问题:

i tried df.groupby("item")["variable"].nunique() and it returns a unique count of every item object.

i want to filter to only return the count of "variable" > 3 conditional on Groupby item... is there a method?

回答1:

When you want the groupby to be mapped to every row of the input, think about transform:

df = df[df.groupby("item")["variable"].transform('nunique') > 3]