Pandas groupby -> aggregate - function of two columns-CodePudding

I'm using pandas aggregate as folows:

In [6]: gb = df.groupby(['col1', 'col2'])
   ...: counts = gb.size().to_frame(name='counts')
   ...: (counts
   ...:  .join(gb.agg({'col3': 'mean'}).rename(columns={'col3': 'col3_mean'}))
   ...:  .join(gb.agg({'col4': 'median'}).rename(columns={'col4': 'col4_median'}))
   ...:  .join(gb.agg({'col4': 'min'}).rename(columns={'col4': 'col4_min'}))
   ...:  .reset_index()
   ...: )

How can I add one more column which will contain sum of values col3 * col4?

CodePudding user response：

First create column new before groupby and then aggregate sum, your solution rewritten in named aggregation is:

counts = (df.assign(new = df['col3'] * df['col4'])
            .groupby(['col1', 'col2'], as_index=False)
            .agg(counts=('col1','size'), 
                 col3_mean=('col3','mean'), 
                 col4_median=('col4','median'), 
                 col4_min=('col4','min'), 
                 both_sum=('new','sum')))