Re: The use of COUNT(*)

From: doid <doid_at_usa.net>
Date: Fri, 10 Jul 1998 18:04:00 GMT
Message-ID: <35A658AC.861A5E01@usa.net>

Been there:)

1. count(*) won't count nulls. this can screw up your logic sometimes...anyway, it's much more efficent to count( <some indexed column> ) or count(1)
2. you don't really want to count every row if you only care about finding the first occurrence of a condition
3. i like: BEGIN

SELECT 1
INTO temp
FROM mytab
WHERE ROWNUM <=1
AND < condition >;

found := TRUE;

EXCEPTION
    WHEN NO_DATA_FOUND THEN
        found := FALSE;
END; 4) alternatively

found := FALSE;
FOR csr IN ( <query> ) LOOP

    found := TRUE;
    EXIT;
END LOOP; It's probably more a matter of style than anything. I think they perform similarly. You should confirm this yourself with TKPROF, or whichever tool you like.

Hope this helped,

doid

Lai King Leung wrote:

> Could anyone tell me about the use of COUNT(*)?
Received on Fri Jul 10 1998 - 13:04:00 CDT