Hi All,
I have a question. I am working on Datastage 7.5.1 parallel extender.
suppose I have to extract from DB2 using parallel extender.
and i do i have to load into DB2 .
1) what are the things i do have to keep in mind while doing the extracting and loading using datastage parallel extender if the source is DB2 UDB and targets are DB2 UDB
2) what are the things i do have to keep in mind while doing the extracting and loading using datastage parallel extender if the source is DB2 Mainframe and targets are DB2 Mainframe
thanks
pavi
DB2 UDB versus DB2 Mainframe
Moderators: chulett, rschirm, roy
-
pavan_test
- Premium Member
- Posts: 263
- Joined: Fri Sep 23, 2005 6:49 am
One thing to keep in mind is parallelism. If you are going against DB2 UDB, then you can do partitioned queries and loads to and from DataStage PX and take full advantage of the parallelism. I am not so sure about the mainframe DB2 option.
I don't know about other users, but we have not had the same experience going against DB2 on the mainframe. While the data may be partitioned on the mainframe, we are stuck with a single stream of data going into DataStage. If we want a 'parallel' query to pull data from DB2 mainframe into DataStage, we ended up creating mutliple queries that run simultaneously pulling subsets of the data. For example, query 1 pulls customers 1 - 1million, query 2 pulls customers 1 million + 1 - 2 million, query 3 pull customers 2 million + 1 - 3 million, etc.
If this is done on DB2 UDB, you write 1 query that returns multiple streams of data, depending on how many partitions you have on your database. If you have 12 partitions, then the output will be 12 streams of data into DataStage.
HTH,
Brad.
I don't know about other users, but we have not had the same experience going against DB2 on the mainframe. While the data may be partitioned on the mainframe, we are stuck with a single stream of data going into DataStage. If we want a 'parallel' query to pull data from DB2 mainframe into DataStage, we ended up creating mutliple queries that run simultaneously pulling subsets of the data. For example, query 1 pulls customers 1 - 1million, query 2 pulls customers 1 million + 1 - 2 million, query 3 pull customers 2 million + 1 - 3 million, etc.
If this is done on DB2 UDB, you write 1 query that returns multiple streams of data, depending on how many partitions you have on your database. If you have 12 partitions, then the output will be 12 streams of data into DataStage.
HTH,
Brad.