Re: Need a Fast Row Count

When you do a count without a where clause Oracle in almost all cases will do a full table scan. It will read all rows and all columns (except long's) no matter what you specify. In some cases adding a hint to use

select /*+ index_ffs(a pk_a) */ count(pk_column) from a;

can sometimes make the count faster.

--
rfazio_at_home.com
http://24.8.218.197/

<kskasi_at_hotmail.com> wrote in message news:38E291C3.CF3DC237_at_hotmail.com...


> Hi Simone

>

> For your info Count(*) is faster than count(1). Unbeleivable Isn't it. But

> thats the fact.

> ...kasi

>

> Simone Jesse wrote:

>

> > How much faster is your count if you use

> > select count(1) from... ?

> > In this case only the first column of every row will be read, not the

 whole


> > row!

> >

> > This might not be as quick as some of the other ideas, but I would be

> > interested in the statistics.

> > Unfortunately none of my tables is big enough to really make a

 difference!


> >

> > Simone

>