Definition
Rows that exist in one dataset (e.g., production) but are absent from another (e.g., the expected post-release state), identified by a key column (primary key or unique index). Keyed unmatched rows are a signal of data inconsistency — a customer record in production with no corresponding row in a reporting table, or vice versa. They often indicate a bug introduced by a recent change.
Why it matters
Keyed unmatched rows are one of the clearest signals of a data bug. A billing record in production that doesn't match the expected schema after a migration can cause errors that are nearly impossible to debug after the fact. Catching keyed unmatched before release prevents customer-facing errors.
How Well Tested handles it
Well Tested's table diff includes keyed unmatched detection: for each table, it compares keyed row sets between environments and reports which rows are present in one but not the other. This feeds into the release risk score and surfaces as a specific reason when risk is elevated.
Related terms
No commitment. 20 minutes. We review your release workflow and show you what's possible.