| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Lincoln Yeoh <lyeoh(at)pop(dot)jaring(dot)my> |
| Cc: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Database I/O and other performance questions. |
| Date: | 2000-12-23 17:06:15 |
| Message-ID: | 24829.977591175@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Lincoln Yeoh <lyeoh(at)pop(dot)jaring(dot)my> writes:
> would/should select count(*) be slower than select
> count(averysmallcolumnmaybeboolean)
There is actually a semantic difference there, see
http://www.postgresql.org/devel-corner/docs/postgres/sql-expressions.htm#SYNTAX-AGGREGATES
count(*) means the count of selected rows, but count(foo) means the
number of selected rows where the variable or expression is not NULL.
count(*) should be a fraction faster, because it doesn't expend any
cycles to check whether the specific column is NULL. The "*" in this
syntax doesn't mean "all columns" the way it does in a select list,
it's just a dummy.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2000-12-23 17:13:31 | Re: Performance and doing USENET style threaded messages |
| Previous Message | Bruce Momjian | 2000-12-23 14:23:49 | Re: Database I/O and other performance questions. |