How to groupby without reset_index() based on multtindex?-CodePudding

data struncture

               col1  col2
A  2021-01-01
A  2022-01-01
B  2021-01-01
B  2022-01-01

This is a dataframe with multiindex(ts_code, date).

Goal

I want to get min date for ts_code. So I have to run df.reset_index().groupby('ts_code')['date'].min(). Is it any method not to reset index to achieve it?

CodePudding user response：

You can either rename your axis for your code to work, or you can pass level=0 into groupby()

Option 1:

df.rename_axis(['ts_code','date']).groupby('ts code')

Option 2:

df.groupby(level=0)

CodePudding user response：

You can use pandas.MultiIndex.get_level_values

df.groupby(df.index.get_level_values('ts_code'))