Backfill by Schema for Fanned-In Tables
Backfills of fanned-in tables can target specific source schemas instead of backfilling every schema that feeds the destination table.
When tables are fanned into a single destination table from multiple schemas, backfills used to be all-or-nothing. That worked fine for most teams - but sometimes you only need a slice. Maybe one schema changed while the others stayed the same. Until now, you'd have to backfill everything, even if just one schema needed attention.
You can now target specific schemas when backfilling a fanned-in table. For example, if a destination table is built from schemas A, B, and C, but only schema C requires a refresh, you can backfill just schema C without touching A or B. This gives you more control, avoids unnecessary data churn, and keeps things lean.
Why this matters:
- More precise backfills - update only what you need
- Faster recovery from schema-specific changes
- Lower compute and warehouse costs by skipping redundant data
- Less operational risk when dealing with large fan-in pipelines