Multiple root_id to id 1:n on web_page_context for structured events


#1

I am busy modelling events of a signup funnel using custom events.

Events have been implemented using JavaScript tracker and are loaded properly to the events table in Redshift. On the modelling side, I intended to apply something close to web-data-model, for struct events instead of page_view ones.

In the scope of that, I discovered that in the table com_snowplowanalytics_snowplow_web_page_1, there are per id multiple root_id == events.event_id. It looks to me that the relation root_id to event_id is 1:1 for page_view events only, but all other events (page_pings, struct) are assigned to the same id. I suppose that these are happening in the same page_view and therefore are bundle under the same id, correct?! (or is not intended)?

If intended, is there another approach to deduplicate the structured events (without partitioning over id, se_category, se_action, se_label and by this approach potentially filtering intended multiple struct events within the same session/page_view (of e.g. a Single Site Application)?

Thanks for your feedback,
Matthias