I have a log table and I need to match up the user_id with potential PRE log-in user_ids
user_id | page_referer | event_time
1199 | blah.com/sign_in | 01/02/2015 1:01:33 am
2244 | blah.com/sign_in | 01/22/2015 7:55:33 am
user_id | page_referer | event_time
2200 | blah.com/before_sign_in | 01/02/2015 1:01:25 am
4488 | blah.com/before_sign_in | 01/22/2015 7:55:12 am
4499 | blah.com/before_sign_in | 01/22/2015 7:55:15 am
so 1199 should matchup with 2200 and 2244 match up with 4488 & 4499 in a result like this:
user_id | other user_id (MV field ok)
1199 | 2200
2244 | 4488, 4499
Part of the problem I'm having is how to construct a subsearch, or join (or appendcols, etc) where I need to use the event_timestamp as a search ( event_timestamp-90 as the lower range and event_timestamp as the upper range).
I could end up with the final result table, or some other join/transaction that can group these pre/post login user_ids.
... View more