Multinode Processing

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
amitkav
Participant
Posts: 3
Joined: Wed Jan 24, 2007 2:39 am

Multinode Processing

Post by amitkav »

If we are using multi node configuration file in DS job, for Join Stage shouldn't we partition the incoming links in HASH mode and on the same key ?

Also if Join Stage has all the links coming from Database tables then from perfomance point of view is it a better idea to use Database Stage with Query having all the required tables with proper SQL joins instead of Join Stage ?
gbusson
Participant
Posts: 98
Joined: Fri Oct 07, 2005 2:50 am
Location: France
Contact:

Post by gbusson »

1- Yes, hash and partition is a good practice. modulus is also useful if the key is an entire.

2- Try the 2 solutions. If your database is powerful and your indexes well set , you can make your SQL query.
Post Reply