基于 Pandas 中的两个第一列合并具有不同行数的列

Question

I have two different files containing the same number of columns but different lengths, ie,

file1.txt

1650,A,1,1,1
1650,A,1,1,1
1650,A,1,1,1
1650,B,2,2,2
1650,B,2,2,2
1650,B,2,2,2
1650,B,2,2,2
1650,B,2,2,2

file2.txt

1650,A,3,3,3
1650,A,3,3,3
1650,A,3,3,3
1650,A,3,3,3
1650,A,3,3,3
1650,B,4,4,4
1650,B,4,4,4

I want to concatenate both of them using pandas such that the result is as follows:

1650,A,1,1,1,3,3,3
1650,A,1,1,1,3,3,3
1650,A,1,1,1,3,3,3
1650,A,NaN,NaN,NaN,3,3,3
1650,A,NaN,NaN,NaN,3,3,3
1650,B,2,2,2,4,4,4
1650,B,2,2,2,4,4,4
1650,B,2,2,2,NaN,NaN,NaN
1650,B,2,2,2,NaN,NaN,NaN
1650,B,2,2,2,NaN,NaN,NaN

I use the following codes but it seems it does not work properly:

df1 = read_data('file1')
df2 = read_data('file2')
result = pd.merge_ordered(df1,df2, how='outer', on=['a', 'b'])

How to solve this problem?

Answer 1

Use GroupBy.cumcount for counter, so possible merge by merge with add column group :

df1['group'] = df1.groupby(['a', 'b']).cumcount()
df2['group'] = df2.groupby(['a', 'b']).cumcount()
result = pd.merge(df1,df2, how='outer', on=['a', 'b', 'group']).drop('group', axis=1)