Daniel Noll wrote:
>
> COUNT() is constant performance if you count all the rows in a table, but
> linear performance if you have a condition on which rows should be counted.
Can you can provide a simple example to demonstrate the behavior that
you are seeing? I tried playing around with some simple (nonsense)
tables but I wasn't able to come up with anything that sounds like what
you are describing.
Are you simply referring to something like:
select count(*) from jobitems;
vs
select count(*) from jobitems where jobid = 4288;
select count(*) from jobitems where jobid > 1000;
A working example might help to clarify what behavior you are seeing...
Army
|