Calculate the difference values with conditions-CodePudding

I want to calculate the difference with conditions, how to do it?

Source date

    TimeStamp  Input  match   
 1      38563  Touch  False        
 2      38568    Pen  False        
 3      38577  Touch  False        
 4      38580    Pen  False        
 5      38589    Pen   True        
 6      38591  Touch  False

expect result

     TimeStamp  Input  match  diff 
 1      38563  Touch  False    NaN     
 2      38568    Pen  False    NaN    
 3      38577  Touch  False    NaN   
 4      38580    Pen  False    NaN    
 5      38589    Pen   True      1    
 6      38591  Touch  False    NaN

CodePudding user response：

Create groups by consecutive Trues and pass to DataFrameGroupBy.diff:

df['diff'] = df.groupby((~df['match']).cumsum()[df['match']])['TimeStamp'].diff()
print (df)
   TimeStamp  match  diff
0      38269  False   NaN
1      38340   True   NaN
2      38341   True   1.0
3      38349   True   8.0
4      38350   True   1.0
5      38351  False   NaN
6      38358  False   NaN
7      38358   True   NaN
8      38366  False   NaN
9      38370  False   NaN