Question About Datastage extract

[palak08]

Hi All,

I am running a server job which has source as Oracle and target as oracle.
Our Oracle and Datastage are installed on separate servers.

I have one SQL query which perform SELECT in 40 seconds. It Selects 20 million records. Now when I do same thing on datastage (select only--> by putting @FALSE condition in transformer) then it takes 35 minutes.

My design looks like
OCI-----> Transformer(@FALSE)--->OCI

In performance stats I can see total record count on input link as 20 million.

My question is when I put @FALSE condition in datastage job and performs only extract part, then what is happing exactly.

Is datastage Extracting entire 20 million records from oracle server and bringing to datastage server?
OR
It is only Selecting on database server?

Any help will be very much appreciated.

[chulett]

It's important to understand the topology here. All selects from sources are brought into the DataStage engine where transformations occur before being pushed to the target. So if each of those components are on different servers you could have a lot of information flying across the network that you need to be cognizant of.

[karthi_gana]

where can i find more information on datastage engine architecture?

[ray.wurlod]

In the manuals, particularly the Information Server Administration and Installation manuals.