hi,

We have a kind of environment where we might have around 10 million rows and looking for optimization in terms of select and insert.
My Question is,
1. I was searching the db2 forums and seeing different type's of table partitioing available ( Like union all view). What are the kind of table partitioning available? Which one would be the best?.
2. I heared partitioning reduces the performance in terms of inserts and updates. Is it true?.
3. Any other way I can opt for solving the 10 million rows issue other than partitioning.
Pls advise!

Thanks!

Partitioning on DB2 for OS/390 (z/OS) is completely different than on DB2 for Linux, UNIX, and Windows.

Partitioning on DB2 for OS/390 is range partitioning, which is not even available (yet) on DB2 for Linux, UNIX, and Windows (range partitioning will be available in V9).

Partitioning on DB2 for Linux, UNIX, and Windows is primarily for the purposes of parallel processing where it is common for queries to require complete table scans. The partitioning scheme is called hash partitioning.

I don't think you will encounter any performance problems with inserts or updates on DB2 for OS/390. But you should avoid frequent updating of the partitioning key, which would cause the row to have to be moved from one partition to the other.

Check out the DB2 for OS/390 manuals to read about partitioned tablespaces. Make sure you don't confuse it with partitioning available on DB2 for Linux, UNIX, and Windows (which is a different product).

Thanks Marcus. That's perfect.