| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Usenet -> c.d.o.misc -> Re: Oracle 8 Partitioning and RAID 5
A copy of this was sent to Konstantin Kivi <konst_at_sirena.rinet.ru>
(if that email address didn't require changing)
On Thu, 25 Mar 1999 10:50:27 GMT, you wrote:
>tkyte_at_us.oracle.com (Thomas Kyte) writes:
>
>>
>> -- you'll scan 1gig. On the other hand, you may be able to use partition
>> elimination on 10 100meg partitions to reduce the full scan to 100meg.
>>
>> o faster mass insert/update/delete: parallel dml (insert/update/delete) is
>> possible with paritions, not so without them.
>>
>> o faster parallel queries: parallel index range scans are possible with
>> partitions, not so without them.
>>
>> o "sliding window": if your data has a sliding window to it (eg: keep the last N
>> months of sales data online, age the old stuff off to another system or archive
>> it), partitions let you do this pretty easily -- load and indesx the new data
>> into a table, take the oldest partition and convert it into a standalone table,
>> create a new partition at the 'new' end of the table and switch this new empty
>> partition with the newly loaded data.
>
>Is it possible to use partitions with multitable cluster?
>The data is organized in 30 tables having common key
>The each key have very limited amount of data associated with it (2-4K).
>It is very convenient because it reduces IO (especially with indexes),
>but almost impossible to manage.
Cluster objects cannot be partitioned. You may be able to use the manual view based partitioning available with 7.3.3 and up though. You don't get the transparent insert/update/delete (you have to goto the base tables for that) though with this.
Thomas Kyte
tkyte_at_us.oracle.com
Oracle Service Industries
Reston, VA USA
--
http://govt.us.oracle.com/ -- downloadable utilities
 |
 |