Page 1 of 1
need help in job design in PX
Posted: Fri Oct 21, 2011 8:25 pm
by sainath
Hi
I have a parallel job which updates to 6 different tables ,I have around 3Million source records...
What is the best may to design jobs (In single job or multiple jobs)
Also I have target DB2 in Z/os.I want to do bulk load ,but this feature is disabled in datastage ..how can i enable this in datastage
thks
Posted: Fri Oct 21, 2011 10:15 pm
by pandeesh
What do you mean by disabled in datastage?
Posted: Sat Oct 22, 2011 8:47 am
by Mike
Bulk load z/OS DB2 from the mainframe.
On my last project I was targeting z/OS DB2 for the first time in quite a while. My z/OS DB2 DBA told me it was not possible to do a bulk load from our aix server (I thought maybe the new DB2 Connector might have that capability). Based on my prior experiences I had no reason to doubt his assertion. I did parallel inserts for several million rows with very good performance, so I had no reason to continue down the bulk load path.
If you really want to do a bulk load:
1) Prepare a load-ready sequential file on your server and ftp it to your mainframe. Check with your z/OS DB2 DBA or the manuals on format requirements. It used to be that nullable fields required a separate 1-byte null indicator field among other requirements.
2) Code a mainframe job (JCL) to do the bulk load. Your DBA probably has tons of examples.
3) Come up with some mechanism to trigger the mainframe job to get submitted. Your DBA can probably help you out with some options.
Mike