Batch job page: why rows were rejected

A batch import from CoreIQ, from a VSAM dataset or an IMS segment, no longer just says how many rows it lost. The job page has a Why rows were rejected card, between Targets and Steps, that names each cause: what the database refused (a NOT NULL violation, a value too long, a type mismatch, a duplicate key, a numeric overflow), in which table and column, and how many rows it cost. Each cause has a How to fix that says what to change in the mapping.

Causes are grouped by what failed, not by the database's wording, so the same constraint failing on different values is one cause and not fifty. PostgreSQL and Oracle errors are both recognised, Oracle by its ORA- code. A step with more than 50 causes shows the first 50 and says how many more there were, so a partial list never reads as the whole story.

Every rejected row's key is kept, not just a few samples, and each cause can export its keys. That gives you the exact records to look at again. The job history shows a run's main cause under its Skipped badge, so you can see why a run lost rows without opening it.