Hi, Scott Marlowe: You said that " As for processing them in order versus randomly,that's a common problem. " do you know why? how postgres work in this scenario.
On Wed, Feb 20, 2008 at 3:07 PM, Scott Marlowe <[EMAIL PROTECTED]> wrote: > On Feb 19, 2008 9:38 PM, hewei <[EMAIL PROTECTED]> wrote: > > Hi,Every body; > > I have a table contains 100,000 rows, and has a primary key(int). > > Now ,I need to execute sql command like "update .......... where > id=*"(id > > is primary key). > > I expect execute 1200-1600 sqlcommands per second(1200-1600/s). > > In test,when the id increase by degrees in sqlcommands, then I can > reach > > the speed(1600/s); > > But in fact , the id in sqlcommands is out of rule, then the speed > is > > very slow, just 100/s. > > Assuming that you're updating a non-indexed field, you should really > look at migrating to 8.3 if you haven't already. It's performance on > such issues is reportedly much faster than 8.2. > > As for processing them in order versus randomly, that's a common > problem. right sizing shared_buffers so that all of the table can fit > in ram might help too. As would a caching RAID controller. >
