I have a dataframe:

  id    value
0  a       w
1  a       l
2  a       l
3  a       w
4  a       w
5  a       w
6  a       l

when I do df.groupby("id").cumcount() it returns:

0  0       
1  1       
2  2       
3  3       
4  4       
5  5       
6  6

I want to count only those ones that equal w in column value and it must be in dataframe form:

0  0              
3  0       
4  1       
5  2       

How to do that with cumcount function?

CodePudding user response：

Use:

res = df.groupby((df["value"] != df["value"].shift()).cumsum()).cumcount()
res = res[df["value"].eq("w")]
print(res)

Output

0    0
3    0
4    1
5    2
dtype: int64

As an alternative:

s = (df["value"] != df["value"].shift()).cumsum()
res = s.groupby(s).cumcount()
res = res[df["value"].eq("w")]
print(res)