Josh Berkus <josh(at)agliodbs(dot)com> writes:
> One of the Sun benchmarking guys ran across this issue, using a fairly
> recent (within the last 4 weeks) 8.3 snapshot. It appears to me that the
> 8.3 planner is unable to tell that, if the value for the first column of
> an index scan is a constant, the second column doesn't need to be sorted.
Works for me. Care to provide a self-contained test case?
regards, tom lane