slow large table

Depending on your access, if you are using a function on the date field and the date field is part of the index, the Oracle may be turning the index off and using a full table scan. If this is true, It may be partially salvageable by making sure the date is the third field of the key in the index, allowing at least range scans.

]In article <6ign2o$oc7_at_bgtnsc03.worldnet.att.net>#1/1,   b-horton_at_worldnet.att.net wrote:
>
> Larry Schenavar wrote:
> >
> > Hello,
> >
> > I have a table with 4 columns. It contains 8 million records and the key is
> > three of the columns. It is very slow to access. Does anyone have any
> > recommendation to increase the performance. Is indexing the three columns a
> > waste?
> >
> > ==============================
> > Larry Schenavar
> > Infrastructure Specialist /DB
> > Electronic Data Systems, Inc
> >
> > Email: schenavar_at_northstate.net
> > ==============================
> If the PK is 3 of the 4 columns as you say, you already have an index.
> Question the date. Is it really required as part of key? I suspect
> that for each query asking for or using the date column, there is
> date conversion going on for each row just to build the result set.
> Consider a date_id instead. If not doing date calcs, make CHAR?
> may help.
> good luck
> bobh
>

-----== Posted via Deja News, The Leader in Internet Discussion ==----- http://www.dejanews.com/ Now offering spam-free web-based newsreading Received on Mon May 04 1998 - 10:03:37 CDT