Re: How best to recreate a table

Perhaps I should add that what I'm doing now is that I have a copy of the table and I copy it back as needed (create table ... as select * from copy). It takes almost an hour, even when set unrecoverable.

I was hoping to use the OS to uncompress an image of a datafile and somehow trick the backup/restore stuff to graft that into a database as a new tablespace. That ought to take just a few minutes. If Oracle can be persuaded no to be too picky about sequence numbers.

I will accept other answers, too.

• kevin

I wrote:
>
> I have a large table (about a gigabyte) that I created unrecoverable.
> I could do it again, but it takes a long time to create, and I need
> to do this over and over (this is a research project, not normal data).
>
> I thought I might do this with backup and recovery, but I'm completely
> new to this (I have no backups at all: just raw data, because I could
> rebuild from raw data in under a day). However, rebuilding this pesky
> table takes enough time that it's slowing me down.
>
> What's the best (i.e. fastest) way to restore a table or tablespace to
> a known point in time without fooling with or altering the other tables
> or indexes in the database? Right now, I'm thinking of SQL*Loader,
> because it's pretty fast. But I was hoping for an OS solution because
> I can restore from a compressed backup in that case.
>
> I'm having a lot of trouble reading through all the considerations and
> options in the Backup and Recovery manual, so go easy on me please.
>
> -- kevin
>
> --
> Kevin O'Gorman (805) 650-6274 kogorman_at_pacbell.net
> At school: kogorman_at_cs.ucsb.edu
> Permanent e-mail forwarder: Kevin.O'Gorman.64_at_Alum.Dartmouth.org

--
Kevin O'Gorman (805) 650-6274 kogorman_at_pacbell.net At school: kogorman_at_cs.ucsb.edu
Permanent e-mail forwarder: Kevin.O'Gorman.64_at_Alum.Dartmouth.org Received on Wed Dec 22 1999 - 23:29:14 CST