no, the results of count(*) and count(1) have always been the same: it returns the number of rows in the result set, regardless the column values all being NULL or not. You can look at it from this perspective:
select t.* , 1 from your_table t;
So you add an extra column to the result table, containing the literal 1 for *every* row. the count(1) function returns the number of non NULL values in that column ...
-- --Original Message-- -- From: Looney, Jason [mailto:Jason.Looney@(protected)] Sent: Tuesday, February 15, 2005 18:17 To: 'lex.de.haan@(protected)'; thomas.mercadante@(protected); joelgarry@(protected); oracle-l@(protected) Subject: RE: is it possible in pl/sql?
My understanding was in the past count(*) returned a count of all non-null rows, where count(1) (or any constant) returned a count of all rows. I just verified this with 10g and this is not the case. The performance difference was to perform a full table scan, instead of counting leaf blocks in a primary key index or something like that. So maybe this was true in 6 or 7?
SQL*Plus: Release 10.1.0.2.0 - Production on Tue Feb 15 10:12:37 2005
Copyright (c) 1982, 2004, Oracle. All rights reserved.
Connected to: Oracle Database 10g Enterprise Edition Release 10.1.0.2.0 - Production With the Partitioning, OLAP and Data Mining options
system@(protected)> drop table test;
Table dropped.
system@(protected)> CREATE TABLE test (t1 VARCHAR2(25) NULL, t2 NUMBER NULL);
Table created.
system@(protected)> system@(protected)> INSERT INTO test values (NULL, NULL) 2 /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> /
1 row created.
system@(protected)> select count(*) from test;
COUNT(*) -- ---- -- 9
system@(protected)> select count(1) from test;
COUNT(1) -- ---- -- 9
system@(protected)> select count(t1) from test;
COUNT(T1) -- ---- -- 0
-- --Original Message-- -- From: oracle-l-bounce@(protected) [mailto:oracle-l-bounce@(protected)] On Behalf Of Lex de Haan Sent: Tuesday, February 15, 2005 6:01 AM To: thomas.mercadante@(protected); joelgarry@(protected); oracle-l@(protected) Subject: RE: is it possible in pl/sql?
no, certainly not. count(*) and count(<any constant>) should return the same result, under all circumstances.there used to be a performance difference, in the past. kind regards,
Lex.
-- ---- ---- ---- ---- ---- ---- ---- ---- -- Visit my website at http://www.naturaljoin.nl -- ---- ---- ---- ---- ---- ---- ---- ---- -- -- --Original Message-- -- From: oracle-l-bounce@(protected) [mailto:oracle-l-bounce@(protected)] On Behalf Of Mercadante, Thomas F Sent: Tuesday, February 15, 2005 13:58 To: 'joelgarry@(protected)'; oracle-l@(protected) Subject: RE: is it possible in pl/sql?
Does anyone know what the difference between count(*) & count(1) was *supposed* to be? Is there a theoretical functional difference?
-- --Original Message-- -- From: Joel Garry [mailto:joelgarry@(protected)] Sent: Monday, February 14, 2005 7:35 PM To: oracle-l@(protected) Subject: Re: is it possible in pl/sql?
Anthony Wilson wrote:
>it's a myth according to the venerable Tom Kyte. The SQL engine silently rewrites >count(1) to count(*):
Funny, he mentions that is a fact, but I've missed the reference to where the fact is shown. Though I'm inclined to believe anything Tom says, I find this particular one ironic in that he is talking about facts and opinions when he says it. And I'm certainly glad he mentions the differing case of 7.x, as that means I didn't get the myth completely out of thin air or confused it with some other function. Just means that old habits die hard.
