2026-10-23 13:54 UTC
@truls46@mastodon.social No, device_id is not present in osupdate. But the table structure and syntax are perfectly valid. It's actually a semantics problem.
Replies (1)
-
@truls46@mastodon.social 2026-10-23 14:01
@petarov@social.tchncs.de I was thinking it might be the visibility of the column name. Strange things happen if the same column name exists in both tables (outer and sub-query). What about osupdate_id? Apart from such a visibility problem, I can't see anything obvious. What exactly is the difference between the systems? Are the tables involved being updated concurrently while the UPDATE runs? SQL Server's different locking strategy might come into play here.