There are a couple of different techiques for performing an update in the Loader transformation. One uses an SQL-based approach and the other uses a data step. If neither of these approaches are fast enough then the only other option is to override the loader and supply your own code for performing the update. You do this by going to the process tab and selecting "User written". Then either enter the code in metadata or enter the name of the file that contains the code you want to execute.
I'm not terribly familiar with efficiently updating Oracle tables - others may have suggestions. From these suggestions you may also find that your data are not set up optimally - e.g. keys and indexes.
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU]On Behalf Of
Sent: Thursday, August 31, 2006 11:45 PM
Subject: SAS Data Integration Studio Update Performance
In the SAS jobs i am using the load strategy as update. it is working
fine if the data is less.
But the program is showing the status as "Running" if the data is more
than 20thousand resords.
I am updating the data in the oracle database.
Is there any way to faasten the program performance for the huge data.
The infrastructure here is good.
the problem is only when using the UPdate. if i use the Append strategy
in the loader then it is faster.
Please help me out.
Thanks & Regards