Hey team, I’m getting a large client set up with ...
# ask-questions
b
Hey team, I’m getting a large client set up with Snowflake + Segment as a data source. Auto generating legacy metrics finds 100+ metrics (great!) but the issue is that the schema is missing from each SQL (
DATABASE.TABLE
instead of
DATABASE.SCHEMA.TABLE
). Meanwhile fact tables don’t appear to allow for auto generating metrics. I have 3 more schemas to add with 100+ metrics each. It seems that I’ll either have to manually set up each fact table, or manually adjust each legacy metric SQL, both very tedious. Is there something I’m overlooking in either scenario?
h
What happens when you click "View Recommendations" in that second screenshot? That's where we do recommend auto-creating metrics on fact tables.
We normally recommend more if more values show up in our scan of the fact table.
One thing you could try is removing the date filter WHERE clause from that fact table, saving it, and seeing if more suggestions appear and creating the fact metrics. Once they are created, then you could re-add the date filter. I'm not 100% sure this would work, though.
b
Thanks Luke! For posterity: turns out our source, medium columns have different names in the tracks table. Not too sure why the query didn’t fail yesterday, but alas — it works now. The fact table now discovers all our metrics 🙏
🙌 1