db2 load

easwaran · Post by **easwaran** » Thu Jan 02, 2003 12:52 pm

what are the different techniques to extract data from DB2 mainframe database to oracle8i target database using Datastage XE 390.
Thanks in advace.

Edited by - easwaran on 01/02/2003 18:46:08

ray.wurlod · Post by **ray.wurlod** » Thu Jan 02, 2003 3:35 pm

The techniques for accessing DB2 on a mainframe from DataStage are excellently described in the manual for creating mainframe jobs (DS/390 Job Developer's Guide), which is in the docs folder with the remainder of your DataStage client software.

Edited by - ray.wurlod on 01/02/2003 15:36:27

easwaran · Post by **easwaran** » Thu Jan 02, 2003 4:32 pm

The techniques for accessing DB2 on a mainframe from DataStage are excellently described in the manual for creating mainframe jobs (DS/390 Job Developer's Guide), which is in the docs folder with the remainder of your DataStage client software.

Edited by - ray.wurlod on 01/02/2003 15:36:27

easwaran · Post by **easwaran** » Thu Jan 02, 2003 4:46 pm

Ray, Thanks for the reply.
What I do not understand from document is
--document explains create the mainframe jobs, generate JCL, upload to mainframe, compile and run at mainframe. Get the out put as flat file delimited / or/ DB2 Ready to load file format.
When other ETL tools give plugin to directly connect to mainframe DB2 database ( just like OCI8 plugin or ODBC stage ) and use as a source in the (server)job( example informatica mapping), connect this source to transformations ,aggregators and finally connect to target table. run the job in the server manager of the ETL Tool(eg Informatica allows like this).
During the job execution, the connection is made to DB2 database and gets the data and transforms and loads to target (oracle) database directly. There is no need to convert the db2 data into flat files and then load to oracle target tables.During job execution both DB2 and Oracle database are up and running.
How this is handled in DS XE390.
Please help e to understand.

Thanks.

ray.wurlod · Post by **ray.wurlod** » Fri Jan 03, 2003 12:43 am

I will have to research this, and don't have access to a DS/390 system for a couple of weeks. If anyone else would like to post a reply, please go for it.

kjanes · Post by **kjanes** » Fri Jan 03, 2003 7:56 am

There is a way to facilitate what you are talking about which is accessing mainframe DB2 from a server side job. If you have a DB2 Connect gateway on an AIX box for instance, that points to a mainframe instance of DB2, you can drop the DB2 stage into your server side design pallete and connect to mainframe data via the DB2 Connect gateway.

Here is my experience... There is a reason why you may not want to do this... Server side functionality is different from mainframe functionality. We ran into some situations where the "server side", while connecting to mainframe DB2, could not handle some traditional mainframe COBOL coding techniques. Like I said, "server" jobs have different functionality from "mainframe" jobs. If you are building hash tables or executing simple ETL from the server side while accessing mainframe DB2 it is probably OK but if you are coding complex ETL, I would code it for the mainframe and not worry about the FTP. FTP'ing data is usally not a big deal and should not be the basis of what platform you build your solution for.

Kevin Janes