Re: index creation fails because temp tablespace is too small

Thanks Lewis.
 As for the build index parallel for local index on partitioned table, I remember oracle does not do necessary use one PQ slave to read one partition, sort one partition, and then build the index for one partition. THat is, it does not split the workload to different PQ slave according to partition range, it split the work load according to the rowid, is it right? I don't find any document talking about this. Can you give document/test result about this? Maybe we have to trace pq coordinator to find how actually oracle does?

So if the above theory is correct, using partition to build the index won't save temp space, unless we use the trick you mentioned above. Right?

But the way, is there any method to let me know that someone else has replied to the thread that I also participate? That will greatly add our efficiency browsing the threads we are interested in.

THanks
CHao Received on Thu Apr 28 2005 - 03:37:17 CDT