We are using v12.01.3942.
Today we encountered a subqquery on a table with 15+ million rows that has been running very well for many years but is very suddenly running very slowly. I fixed by forcing index usage by adding "WITH INDEX(index_name)" clause.
My question is why all of the sudden would this change - I think it happened mid-day and is there any preventative measures to keep thinks working reliably. Dropping/re-creating statistics or any other ideas.
asked 25 Nov '13, 17:58
Try to analyse query plan, and you find, why server choose another strategy. May be will be better, if after serious data updates, you will run UPDATE STATISTICS instead of index hinting.
answered 26 Nov '13, 17:31