How to sum subset of columns when meet certain number(condition)

Question

I have a dataframe which is shown in the picture

I want to add the value of different subsets of non-zero 'ttr' and get the value of 'hour', respectively.

The results should be looked like: hour:7, ttr:15seconds, hour:7, ttr:25seconds, hour:8, ttr:15seconds, and etc.

The difficult part is how to add different subsets of 'ttr' column when meet zero seconds.

Does anybody know how to solve this problem? Thank you!

Answer 1

IIUC filter your df , then using the index to get the continue hour into one group

s=df[df.hour!=0]
s=s.reset_index()
s.groupby([s.hour,s['index'].diff().ne(1).cumsum()]).ttr.sum()
Out[389]: 
hour  index
7     1       00:00:15
      2       00:00:25
Name: ttr, dtype: timedelta64[ns]

---

Data input

df=pd.DataFrame({'hour':[0,0,7,0,7,7,7],'ttr':[0,0,'00:00:15',0,'00:00:05','00:00:16','00:00:04']})
df.ttr=pd.to_timedelta(df.ttr)