On 04 April 2014 18:09, Joshua Yanovski Wrote:
> The counter would be updated only by VACUUM, which would perform the
> same computation performed by the COUNT operation but add it
> permanently to counter just before it set the visible_map bit to 1 (I
> am not certain whether this would require unusual changes to WAL
> replay or not).
I think there is one more disadvantages with this (or maybe I have missed something): User may not use count(*) or may
usejust once after creating new kind of index proposed but VACUUM will have to keep performing operation equivalent to
iterativecount(*), which might degrade performance for VACUUM.
Thanks and Regards,
Kumar Rajeev Rastogi