CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
leo
leo

Reputation: 441

pandas groupby with multiple groupings

I am trying to groupby 2 columns with an OR condition to create a group categorical count where if a number is repeated in either groupid1 or groupid2 it will create a 3rd unique grouping number.

input df

ID      GROUPID1    GROUPID2    
F632    173         1315    
F632    173         1315    
F966    173         1519    
F966    134         1519    
F173    704         1437    
F455    704         1201    
F456    702         1202    
F457    800         1500    
F966    173         1519
F966    134         1519

desired result - where 173/134 or 1315/1519 create groupid3 '1'; 704 = 2, and 702 and 800 are unique.

ID      GROUP1  GROUP2  GROUP3
F632    173     1315    1
F632    173     1315    1
F966    173     1519    1
F966    134     1519    1
F173    704     1437    2
F455    704     1201    2
F456    702     1202    3
F457    800     1500    4
F966    173     1519    1
F966    134     1519    1

Upvotes: 1

Views: 59

Answers (1)

BENY
BENY

Reputation: 323386

Let us try duplicated with or

s=(~(df.GROUPID1.duplicated()|df.GROUPID2.duplicated())).cumsum()
0    1
1    1
2    1
3    1
4    2
5    2
6    3
7    4
dtype: int64
df['GROUP3']=s

Upvotes: 2

Related Questions