How to drop patterns-CodePudding

df:
Id activity  sequence timestamp
1   start       1      2020-06-12 09:51:42
1   end         2      2020-06-12 09:51:42
1   start       1      2020-06-12 09:58:52  
1   end         2      2020-06-12 10:12:22

I wanted to drop the middle part of the repeating process and only get the the first and last timestamp.

this is the output i hoped for:

df:
Id activity  sequence timestamp
1   start       1      2020-06-12 09:51:42  
1   end         2      2020-06-12 09:58:52

Thanks in advance

CodePudding user response：

Try .loc update:

df.loc[df['activity']=='end','timestamp'] = df.loc[df['activity']=='start', 'timestamp'].values

CodePudding user response：

I think you are trying to group by Id, activity, sequence et get minimum date or maximum, and not drop duplicates

Go split your df in start, and end, and group by to get minimum date for start et get maximum for end, and then concat results