Question : Select count(*) query

I need to retrieve the number of rows from syscolumns table
which wd give me the faster result?
select count(*) from syscolumns where object_name(id) = 'employee'
or
select count(1) from syscolumns where object_name(id) = 'employee'

and why?

There is another way using exec sp_spaceused <'tablename'>
Can we retrieve the number of rows in any way by executing this statement?
Is there any way of improving this query?

Thanks,
rose

Answer : Select count(*) query

Rose,

The reason to use count(1) instead of count(*) is simple.
If you use count(*), then you are instructing sybase to fetch the complete row of table ABC.  If your table has 20 columns totalling 1024K, then sybase must fetch all these bytes (note they are NOT returned to your query but they are touched in the process).  If you use count(1) then sybase is just "handling" the 1 byte.  This is usually a minor P&T issue, but why not start coding all your queries this way.  All those IOs can add up over time.

As far as the sp_spaceused, you might want to go to
www.edbarlow.com then look at his SPs (specifically sp__helptable).

In general
If you find a sybase SP that you like, then the thing to to is to create
a modified version of this SP and save it as "sp__xxxxx".  Basically, the double underscore indicates a "homegrown" or internal SP.

hope this helps
knel