Re: Alter Table on 41 million rows (2G):

sujata_soman_at_omm-la2-infotech-001@internet.omm.com wrote: > > Hi all, > Background Info : Informix ver 7.22 UC4, HP-UX 10.20. > I need to alter a table which has about 41 million rows ( 2Gig data ), > basically changing the data type of one single column. This table has 40 > fragments ( don't ask me why ) > My question is as follows:- > a. What are the factors that influence the performance of this query ? Amount of available temp space, load on the system, I/O traffic, and such. > b. What do I to speed up the process ? Since it's only 2GB of data, you are better off unloading the data using ipload, recreate the table with the column changed, and reload data. This also ensures that the data conversion will not create any kind of performance problems, because the unloaded data is in ascii. > c. How does setting it as a PDQ process compare it to running it as a normal > process with PSORT_NPROCS set to an appropriate value > d. Would running the query generate 40 parallel session ( for 40 fragments > ), or do I have to force it in some way ? > > TIA > > Sujata > > ***************************************** > Sujata Soman > > > E-Mail: ssoman@omm.com > > O'Melveny & Myers, Los Angeles > > ******************************************