Re: count(*) vs count(columnX): Which one is better?

Robert Chung wrote:
>
> Regarding an SQL statement like this:
> SELECT count(*) FROM tableX where columnX='ABC'
> and
> SELECT count(columnX) FROM tableX where columnX='ABC'
>
> Which one is better for performance? What if columnX has an index?
> Would there be some differences in performance and efficiency? Thank
> you in advance.

These should be equivalent. However, if you changes the syntax as follows, you could get different counts provided columnX allows NULL values.

	select count(*) from tableX;	    -- returns all rows
	select count(columnX) from tableX;  -- returns rows with non-null 
					    -- columnX values.

In your example, though, the results SHOULD be identical.

-Bill Boyle Received on Sat Oct 24 1998 - 23:04:17 CDT