Re: DB2 HADB

Mark A apparently said,on my timestamp of 26/11/2005 12:43 PM:

> clustering is used more often. On DB2 for z/OS (which you claim to know
> about) clustering is the default even if not explicitly defined (using the
> sequence of the first index created), so pretty close to 100% of tables in
> DB2 for z/OS use table clustering.

Obviously! The thing uses VSAM under the covers to keep the data and the only way to index VSAM efficiently is to use the multi-key ISAM. Which is exactly what clustered tables are in DB2 for z/OS! Come on, Mark: both I and DA have used MVS/XA and VM all the way back to DOS/VSE, don't try to hide the obvious! Besides: I have access to DB2 for z/OS doco, it explains how VSAM is used. ;)

What I find suprising is that in db24luw this is also still the case. I wonder: did IBM port VSAM to those environments "under the covers"? Not that it'd be a bad thing, mind you: VSAM would certainly be FAR superior to most native file system handling in those platforms. It's one of the few mainframe things I sorely miss.

-- 
Cheers
Nuno Souto
in rainy Sydney, Australia
wizofoz2k_at_yahoo.com.au.nospam