three items:
- specify the storage parameters, derive the values very carefully (I use
data from user_tables after analyze table compute stats)
- use the parallel option e.g.,
create index I_my_table on my_table (my_field)
storage (initial next etc)
PARALLEL (degree 4) -- I used 4 here but you can use up
to 4 * number of cpus, tho for multi-cpu systems, 4 is good.
/
- create your indexes with multiple concurrent sessions. This works
particularly well if large amount of cache (db_block_buffers) and indexes
are all on same table
"Andreas Wizemann" <AWizemann_at_fvvag.de> wrote in message
news:0745vts4r261d76qckfd3vobb5bdkrgm8d_at_4ax.com...
> Hi,
> i need to create multiple indexes on very large tables.
> Currently i do this one by one, which takes a lot of time.
> How can i speed up this process ?
> Is there a way to do something in parallel ?
>
> Regards
> Andreas
> Andreas Wizemann
Received on Fri Nov 16 2001 - 14:06:06 CST
