Re: Rtree index creation very slow

Jose Luis Matute wrote: > It doesn't parallel (only one CPU used, although PDQ > parameters are set 'correctly'). > I think the rtree is written using the Virtual Index Interface. If this is the case then PDQ will not help because a secondary access method created using the VII cannot take advantage of the parallelism. An IBM Global Services guy working for a large UK retailer once told me that a btree is created using multiple read & sort threads, a merge thread, and multiple loader threads. I found empirically that when creating an index using a VII blade, there would only be one thread, even when the UDRs were "parallelizable" and there were multiple table fragments. Cheers, Donny