Re: Why percent_rank is so slower than rank?

Tom Lane <tgl@sss.pgh.pa.us> wrote:
 
> Do you have reason to think that anybody is likely to exercise
> window functions in HEAD, beyond what the regression tests do, in
> the next couple of months?
 
Not specifically, no.  From the description (not having read the
patch) I was somewhat concerned that it might affect something
outside that narrow use case.  If that's not possible, then I'm more
comfortable putting it in a release that soon after it hits the
repository.
 
It's a judgment call, and you're clearly in the best position to
make it.  You asked for thoughts, so I shared my concerns.  :-)
 
-Kevin