Thanks guys,
I asked this as we are making our tables storage / transaction stuff configureable by the clients, through the product, kinda moving some of the frequent DBA stuff into the application. While the product gets delivered with standard settings and they would normally start with these settings they may realise 6 months down the line that our initial "guesstimates" are not accurate for there data loads. These are pretty much the only changeable things for a table (once it has been populated) as I've been reliably informed that you can not alter the initial extent size of the table once it's been created (only drop and recreate with new size). My concern is that if the client changes the pctfree/used while there is data in it it will only effect changed data from that point on and so any DBA worth his salt would want to import export the data so that it all obeys the new rules. Is this the case??
If so, rather than do an actual import export of the data I would rather do a "create table as select" and then an "insert into select" instead as this can all be done within PL/SQL and give me greater error control. I was also thinking of doing the initial "create as select" with the no logging option as some of these table contain 5 - 10 Million rows and with the logging on the transaction could take forever. How dangerous is this though? would it be best to ensure a resent backup has been done incase there is some system crash while the operation is completed? Would this normally be done by a DBA anyway?
Thanks again for the valuable, not to mention free advice you all give.
Mike.