Re: Confused about Paritioned Tables

bbristol_at_ix.netcom.com wrote:
> I'm aware that each partition can be in its own tablespace, but does
> Oracle actually create TABLES with the partition name?
>
> i.e. If you create table EMPLOYEES with partitions PART1, PART2, and
> PART3, are tables created with those names?

You create a table X with partitions P1 P2 P3 etc. you do not create tables P1 P2 P3... at least not on the user level. oracle will create 'objects' with those names :) thats the difference. You DO create or Oracle does create seperate segments. but you cannot just insert into the partition without adding or specifying the table name.. and the reason they can be in seperate tablespaces now is, because the oracle engine just sees every partition as a seperate Data object, we now have object ids and dataobject ids.. if you look at a partitioned table, it has an object id, and all partitions have a dataobject id because it actually has a physical segment.

you can do on a partition what you can do with tables, execpt for a few restrictions. a small issue here could be that is updating the partition key, you cannot update a row so that it would move to another partition, at least not in 8.0.

Main reason to use partitions like we have them in 8.0 would be historical data and maintenance. Or maybe in a parallel server environment or a place where you can easily have applications run on seperate partitions (we only support partition by range right now).

So what is really important here is to get a good usable partition key. This very important.

If you want to discuss this further, please feel free to post or mail me...

cheers
Wim Received on Wed Jan 21 1998 - 00:00:00 CST