This project is read-only.

All NULL attributes missed in Extract


As supplied, rows with all attributes Null will be missed/ignored in the Extract. Maybe this is what you want. And maybe this situation is so unlikely to ever occur, that you can safely disregard it. - I had to extract a bridge/relation table with three keys and just one attribute, a date, which was frequently Null. - The solution was to enter the "Did Record Change" conditional split task and add a new top priority (=first!) condition called "New Row", defined simply as: ISNULL(VersionId_ETL) == TRUE. - Then, immediately after, both New Row and Change data flows are directed to a Union All task and then passed to InitEtlFwk2, as before. - Of course, it is only to keep the overall flow clean and simple, that I choose to first split, then straight away join the new and changed rows.
Come to think of it, it might be a performance improvement to first have a simple test condition for new rows, and only apply the long and windy Change conditions to matched rows.
For purists, it would probably be even better and nicer to use the "Redirect rows to no match output" straight from the Lookup component as input to the later Union All task, instead of resorting to the "Ignore failure" option there.