Lookup direct from Oracle database
Posted: Wed Jun 14, 2006 10:32 pm
When I have a lookup stage with reference link direct to Oracle stage what exactly happens in the background?
Does this simply create a temporay lookup fileset which is then loaded into memory in "entire" mode on all nodes? or Does it create partitioned lookup filesets? Or does it not really create lookup filesets at all?
If a lookup fileset is only required for one job (MainJob) and will be recreated just before the each run of MainJob, will I get the same performance by having a separate job that loads the lookup fileset and then MainJob uses that lookup fileset compared to the easyier implementation of having the lookup stage connected directly to the Oracle stage?
Does this simply create a temporay lookup fileset which is then loaded into memory in "entire" mode on all nodes? or Does it create partitioned lookup filesets? Or does it not really create lookup filesets at all?
If a lookup fileset is only required for one job (MainJob) and will be recreated just before the each run of MainJob, will I get the same performance by having a separate job that loads the lookup fileset and then MainJob uses that lookup fileset compared to the easyier implementation of having the lookup stage connected directly to the Oracle stage?