Checking and filtering the first event of each group in Pandas

Question

I have an event dataset. First of all, I grouped my dataset by user_id and location with pandas library. Then I order my each group by created_at. I want to delete all group if the group does not start with 'Open' event.

My dataset:

user_id	location	event_name	created_at
A1	AAA	Event5	9/1/2020 9:06:52 PM
A1	AAA	Event6	9/1/2020 9:07:02 PM
A1	AAA	Open	9/1/2020 9:07:12 PM
A1	AAA	Event1	9/1/2020 9:07:22 PM
A1	AAA	Event2	9/1/2020 9:07:32 PM
A1	AAA	Event3	9/1/2020 9:07:42 PM
A1	AAA	Event2	9/1/2020 9:07:52 PM
A1	AAA	Open	9/1/2020 9:08:02 PM
A1	AAA	Event1	9/1/2020 9:08:12 PM
A1	AAA	Event2	9/1/2020 9:08:22 PM
A2	AAA	Event9	9/1/2020 9:07:22 PM

I want:

user_id	location	event_name	created_at
A1	AAA	Open	9/1/2020 9:07:12 PM
A1	AAA	Event1	9/1/2020 9:07:22 PM
A1	AAA	Event2	9/1/2020 9:07:32 PM
A1	AAA	Event3	9/1/2020 9:07:42 PM
A1	AAA	Open	9/1/2020 9:08:02 PM
A1	AAA	Event1	9/1/2020 9:08:12 PM
A1	AAA	Event2	9/1/2020 9:08:22 PM

Thank you.

Answer 1

In short:

df[~(df['event_name'] != 'Open').groupby([df['user_id'], df['location']]).cumprod()]

Explanation:

(df['event_name'] != 'Open') gives us a series of True/False which can also be understood as 1/0 when interpreted as numbers. cumprod apply cumulative product to the True/False series so that as soon as it sees an False (i.e. 0 or 'Open'), it returns a zero and remain zero thereafter.

So we may use zero as the indicator of rows that we wanna to keep.