Sales import processor fix
You can now import sales via CSV without hitting processing errors.
We fixed a bug in the background import worker that caused certain CSV files to fail mid-processing. Mid-processing is the worst failure point for an import: some rows in, some rows lost, and no obvious line between them.
Imports that stalled or errored out now run through cleanly. Your mapped fields — including the promotional code column — get processed end to end, so coupon-attributed sales land against the right affiliates.
- CSV files that previously failed partway now complete.
- All mapped fields are processed through the full pipeline.
- The promo code column carries through, keeping coupon-based attribution intact on historical data.
If a past import failed, just run it again — it should complete this time, and the import processor's conflict handling takes care of any rows that made it in during the failed attempt.
Bulk sales import is how historical revenue gets credited when you migrate or backfill, so this path needs to be boring and dependable. It is now.