How to Access Groups From a Groupby in Pandas Dataframes

There is an easy method to get the groups from a groupby operation.

import pandas as pd

df=pd.DataFrame({'A':[1,1,2,2,3],'B':['a','b','a','c','b'],'C':['a','b','c','d','e']})

df

   A  B  C
0  1  a  a
1  1  b  b
2  2  a  c
3  2  c  d
4  3  b  e

Let’s do a groupby and save it to a variable.

group=df.groupby(by=['A'])

At this point we can see the groups running the following:

group.groups

{1: [0, 1], 2: [2, 3], 3: [4]}

The keys of this dictionary are the unique values of Column A which we applied the group by operation. The values are the indexes of the rows where every group has. If we use the indexes, we will get the corresponding group.

#lets get the group 1
df.iloc[groups[1]]

   A  B  C
0  1  a  a
1  1  b  b

However, Pandas has also its own function to get the groups.

group.get_group(1)

   A  B  C
0  1  a  a
1  1  b  b

What if we have more than one group variable? It’s the same as before but we have to use all variables inside a tuple. Let’s see the groups to understand this concept.

group=df.groupby(by=['A','B'])

group.groups

{(1, 'a'): [0], (1, 'b'): [1], (2, 'a'): [2], (2, 'c'): [3], (3, 'b'): [4]}

For example, if we want the group has the value 1 for A and the value “a” for B we should run the following:

group.get_group((1, 'a'))

   A  B  C
0  1  a  a