New: Cookbooks and AI ExplanationsStep-by-Step recipes to solve problems connected to Roadmaps and Cheat Sheets. Need more details? Use AI buttons for structured and simple explanations with concrete examples throughout the whole platform.Take a look
Ten thousand rows load fine until row four thousand has a bad date, and the whole job dies right there.
What you'll have at the end
A quarantine table holding every skipped row along with the reason it failed
You need
A nightly load that already stages tonight's new rows in a table as plain text, a separate target table whose columns carry their real types, and write access to both.
Not covered
Building an aggregate check that stops a whole batch before it starts; this recipe assumes the batch already began and handles what happens when one row inside it turns out bad.
Give every load a data quality gate
go there first for a check that can stop a whole bad batch before it starts; this recipe instead lets every good row through and only sets the bad ones aside.
Keep a script running when one row in the file is bad
reach for this instead when the load itself is a script working through rows one at a time, rather than one set-based database statement.
Checked 25 Aug 2026
Part of the ETL (Extract, Transform, Load) cookbook