Pandas Concatenate Values From Different Rows

Question

Given this dataframe:

import pandas as pd
a=pd.DataFrame({'number':[2,2,3],'A':['abc','def','ghi']})

a
    A   number
0   abc     2
1   def     2
2   ghi     3

I need to concatenate values, in order of index, from rows with the same number value, separated by '; '.

Desired result:

    A           number
0   abc; def    2; 2
2   ghi         3

So far, I thought I could isolate the dataframes and then somehow try to join them together like this:

a['rank']=a.groupby('number').rank()
a1=a.loc[a['rank']==1]
a2=a.loc[a['rank']==2]
b=a1.merge(a2,on='number',how='left')
b=b.fillna('')

b
    A_x     number  rank_x  A_y     rank_y
0   abc     2   1.0     def     2
1   ghi     3   1.0     

..and then it's just a matter of something like this per column:

b['A'] = b['A_x']+'; '+b['A_y']

...but is there a more concise way to do this (perhaps for all columns at once)?

Thanks in advance!

Answer 1

Use groupby + agg -

a.astype(str).groupby(a.number, as_index=False).agg('; '.join)

          A number
0  abc; def   2; 2
1       ghi      3

Thanks to MaxU for the tune-up!

Answer 2

You need a new para to help then groupby + agg+ join

a.assign(number2=a.number).groupby('number2').agg(lambda x : ';'.join(x.astype(str)))
Out[238]: 
               A number
number2                
2        abc;def    2;2
3            ghi      3