Drop all indexes and do the update then create the indexes.
Robert
"Iain Wiseman" <bibble_at_ibm.net> wrote in message
news:3a1f05ef_4_at_news1.prserv.net...
> I would like to update two columns in a table in bulk the fastest way
> possible. In general we do not have any insert or deletes, just updates.
> There are few indexes over this table as well.
>
> We have made the initial extent a quarter of the expected total size. (3m)
>
> What is the quicket way to do this ?
>
> EXEC SQL
> for :rowCount
> UPDATE
> MY_TABLE
> SET
> value = value - :DeltaValue,
> ba.last_use = SYSDATE
> WHERE
> id = :myId;
>
> I am currently using array updates but the CPU usage is very high. Are
there
> ways to create views that contain just the key and the two columns and is
> this efficient ? What are my options. (This is currently in ProC).
>
> We are looking to process over 200+ on an POLO running HPUX 10 something
> and oracle 8.0.6
>
> Thanks
>
>
Received on Sun Nov 26 2000 - 08:48:18 CST
