Re: Migrate 200 Million of rows

On Tue, 22 Jul 2003 03:50:10 +0000, Nuno wrote:

> Hi,
>
> I would like to migrate 200 Million records. Furthermore my new table
> will have more fields that the old one. The new table will be in
> different database therefore a link will be used as well.
>
> I was thinking about:
> Strategy A)
> 1) create table ... as select * from
>
> with a join with temporary table to make the new table with new
> columns
>
> 2) create index, contraints etc
>
> Strategy B)
>
> 1) create table
>
> 2) create index and make it unusable (I have to check if this is
> possible in v8.0.5)
>
> 2) Insert into /*+ APPEND */ select * from
>
> 3) rebuild the index, constraints etc
>
> Could you please help me in the best strategy to perform this task?
>
> Do you have any more suggestions

When you say new table will have more columns , can you add them / update them after the table has been created . I would try to do with less redo / undo as much as possible .
Create a new empty table
Insert into new table with old table with append hint and with degree of parallelism with half as many cpu's as you have Then build indexes with nologging and parallel Then add columns using alter table and update them and .. Post as how you did it if you have already done it and problems encountered . Received on Thu Jul 24 2003 - 19:43:32 CDT