Re: optimizing count statement

[Karolis]

I presume this could be because MySQL is not able to process bigger amounts of data in RAM, so it has to write it to a disk. That would explain why at some point the query time jumps so quickly. I have also noticed that mysql can count more rows quickly if there are no table joins. Say 6000 instead of 3000.

I might refuse to process queries which will generate more than X results (3000 in this case). The problem is, however, that I have no way to presume how many results mysql is going to find. This is a search engine implementation. It can find 1 result or it can find 100 000 results.

Is there a way to tell count(*) function to stop counting at a certain point? Say if it counts 5000 rows, just return 5000 instead of scanning until the end?

thanks,
Karolis