Yes, would be redundant, but the parser understands better the command, takes less time and assures to reserve limited memory results.

This can be checked and tested for a 1 milion record table, selecting conditional to am indexed field, example:

SELECT count(1) FROM cities
WHERE country_id = 50
LIMIT 1

SELECT COUNT(1) FROM users WHERE postal_code = ’23456′

If you have indexes on postal_code, these count takes few time.

Also, take a look that I make count(1), not count(*), because I’m saying to mysql that I don’t need fields, only count, and the query it’s executed quickest.

Remember that the best policy to make a count before to listing results (and paging), on large sites and large tables, it’s to have a general count in a separate table, almost for those tables that you need to make full scan (or limited where). Also, you can obtain full count on base mysql tables and indexes.

In relation to LIMIT, on all queries it’s better to use a LIMIT:

SELECT count(1) FROM users WHERE postal_code=’67677′ LIMIT 1

SELECT count(1) FROM users LIMIT 1

Even when you have 2 or 3 millions of users, these counts will be made on less that a sec.

Remember ever to use cache in your code, and to save those queries that can take a long time.

Any how thank a lot of this better workaround.

SELECT COUNT(1) FROM (SELECT 1 FROM subscriber LIMIT 18, 2000)s

Thanks,

Binod Suman
Bangalore,India