ValueError: Cannot index with multidimensional key after pandas update

user_6689902 0
Viewed 2062

I got this error after upgrading to pandas 1.2.0:

ValueError: Cannot index with multidimensional key

This is my "semplified" code:
data_final = data.loc[data.groupby(by=["field1", "field2"], as_index=False)["field5"].idxmin()][['field1', 'field2', 'field3', 'field4', 'field5']]

This code was working fine before upgrading to pandas 1.2.0

What's wrong now ?

pandas python
1 Answers

First create default unique index and then remove ] after idxmin and as_index=False:

data = data.reset_index(drop=True)
cols = ['field1', 'field2', 'field3', 'field4', 'field5']
data_final = data.loc[data.groupby(by=["field1", "field2"])["field5"].idxmin(), cols]
user_29010020
Related