Question

I have a query that works

$query_list_records = "SELECT COUNT(*)
                         FROM (SELECT comment_id from myl_r_comments
                                WHERE release_id=? LIMIT ? OFFSET ?) A";

This does COUNT all rows from the 2nd SELECT.

But for optimisation purposes,
why does the following line not work?

$query_list_records = "SELECT COUNT(*)
                         FROM myl_r_comments
                        WHERE release_id=? LIMIT ? OFFSET ?";

I figure, for a lot number of rows, returning only the COUNT value is nicer than returning a lot of rows (one column selected) and then counting them.

Let me know what you think.

Answer 1

In the first version, the LIMIT is applied and then the COUNT(*) is taken of the result: e.g. the subquery returns X records, which are limited to Y, and COUNT(*) in the outer query returns Y.

In the second version, the COUNT(*) of all the records is taken and then the LIMIT is applied (but there’s only one, aggregated record to limit): so X is the sole record returned (but the OFFSET might cause it to be excluded from the resultset).

You could instead do:

SELECT LEAST(GREATEST(COUNT(*) - ?, 0), ?) from myl_r_comments WHERE release_id=?