Wsrtjtyk

I am a new pandas user and I would like to ask is it possible to do this?

Here is the sample of my data frame. all data type of both attributes are the string.

id class

A1  X1,41

A1  X1,42

A1  X1,43

A2  X1,41

A2  X1,45 

I have merged the data frame using groupby and the results are being like this. df = df.groupby(['id']).sum()

id class

A1 X1,41X1,42X1,43

A2 X1,41X1,45

but I would like the results to be like this (hope the data contains in the form of list.)

id class

A1 [X1,41], [X1,42], [X1,43]

A2 [X1,41], [X1,45]

I think you are looking for this:

df.groupby('id').apply(lambda x: [[_x] for _x in x['class']])

This means to groupby the 'id' column, and for every grouped object, apply the function given. In this case I provide a function that creates a list of lists of the objects in that groupby object. _x could be anything, I just named it that to reflect that it is temporary and a placeholder.

Expanding on @Ethan Koch's answer:

df.groupby('id').apply(lambda x: [[_x] for _x in x['class']])

returns a Series, not a Dataframe. To convert back to Dataframe:

df2=pd.DataFrame({'id':df.index, 'class':df.values})