![]() These sessions already have a row in ssions, which we will drop just before we update the table. This will happen when a session was still active when the pipeline last ran, or if not all events had arrived yet. There are a couple of important things to note here.įirst, some sessions will have events that were processed in earlier runs. YY% are within the last week so we don't look at older eventsĪND domain_sessionid IN (SELECT id FROM ssion_id ORDER BY 1) in the unstructured event or context tables). In some cases, it’s easier if we can restrict on the event ID, so we don’t need to join to atomic.events to get the session ID (e.g. ![]() INSERT INTO ssion_id (ĪND etl_tstamp IN (SELECT etl_tstamp FROM scratch.etl_tstamps ORDER BY 1) Select all session ID that have at least one event in the batches (or batches) that we want to process. If either the pipeline or the SQL break for some reason, the problem will need to be resolved within one week or some batches that still need to be processed will be excluded (the filter can of course be updated if that were to happen).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |