RE: HPL vs. Select Into

If you use named pipes to unload and load the data in parallel (one onpload writing to a fifo and another onpload loading from the same fifo) you can move a large table pretty quickly. This strategy has the added benefit of eliminating the intermediate step (and bottleneck) of writing an unload file to disk (much like the insert into select * does). You can use multiple named pipes in a device to take advantage of multiple CPUs, modify the onpload.std config parameters and configure a No Conversion job to boost performance. I did this recently with the following results on a 4 CPU machine with disk attached to a SAN: Number of rows: 45 million Row size: 357 bytes (includes a varchar (250)) Unload/reload time: 25 minutes I doubt you can get insert into select * from to outperform this. Andrew Ford -----Original Message----- From: informix-list-bounces@iiug.org [mailto:informix-list-bounces@iiug.org] On Behalf Of Superboer Sent: Tuesday, October 03, 2006 3:12 AM To: informix-list@iiug.org Subject: Re: HPL vs. Select Into HPL in express mode is the fastest. test it you'll see. when you create the target tables make sure the first/next extent size are properly configured. and a checkpoint has been done before loading. Superboer. Tam OShanter schreef: > Hello, > > Looking at moving a table with many millions of rows to a new DB Space. > > Question being, will the move be faster using HPL that doing something like > "Select * from into......."? > > Why, why not? > > Thanks for the advice friends, > > Tam. _______________________________________________ Informix-list mailing list Informix-list@iiug.org http://www.iiug.org/mailman/listinfo/informix-list