Alan McDonald wrote:

>
> PLAN (ERSJOB INDEX (ERSJOB_IDX2,ERSJOB_IDX4))
> PLAN (ERSJOB INDEX (ERSJOB_IDX1,ERSJOB_IDX4))
>
> PLAN (ERSJOB INDEX (ERSJOB_IDX4,ERSJOB_IDX2)) <-- different order
> PLAN (ERSJOB INDEX (ERSJOB_IDX4,ERSJOB_IDX1)) <-- different order
>
> and it's very slow...

This is really weird and I would appreciate a test case. It shouldn't
make any difference what index scan order is chosen. Maybe something
else is a problem?


Dmitry