----- Original Message -----
From: "Paul Novitski" <paul@xxxxxxxxxxxxxxxxxxx>
While we're talking about optimization, I'd want to check to make sure
COUNT(*) didn't ask MySQL to generate a throw-away recordset consisting of
all fields. I wonder if it would be more machine-efficient to use
COUNT(`last_name`), specifying a single field, in this case the same field
your query is examining anyway.
If the fields mentioned in the where clause are keys or if there is no where
clause, a count(*) uses the index trees to make the count, it won't even
access the data racords at all. Moreover, it won't even read the whole
index tree node, just the header record of the block that indicates how many
key records that block contains. I don't really know about MySql or any
database engine, but these are techniques well known in the trade.
On the count(fieldname) issue, it is different from a count(*) and may give
different results. When you mentione a field it will only count records
where that field is not null. Select count(fieldname) from table is the
same as doing a select count(*) from table where fieldname is not null. It
will take just as longer as processing any additional where clause.
Satyam
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
