DSXchange

Folks,

We are migrating our DataStage from Windows to Unix and have hashed files that contain Surrogate Keys.

If we can replicate the Hashed File directory structure under Windows to a corresponding structure in the target environment (AIX) and migrate all of the files/Data found in that structure, however I am not convinced . . . will this produce the desired results??.

What are the other things we need to make sure are done for our persisted Hashed files to be in the same state after the move??.

Thanks in advance for your ideas.

Bryceson

I also had similar questions, take a look at the following post
viewtopic.php?p=182232

They won't be in the same state. You're going to need to dump your hashed files to sequential text files, ftp them to the unix box, then load them into the hashed files on the unix side. This will cause your hashed files to build fresh.

If you have STATIC or specialty tuned (minimum modulos, etc) hashed files, you're going to need to create them the exact same way first on the unix side.

Ken,

Thanks for your knowledge!!

How do I need to handle the surrogate key that are created in DataStage at the project level then if the Hashed files are going to build fresh??

Thanks . . . Bryceson

If you are talking about your hashed files created within the DataStage projects, you must also dump them to files and create on the unix side. If you mean things like the hashed files used in the SDK functions, you must copy those over as well.

The thing to consider on those files is to build a job that refreshes them anyway by doing SELECT MAX(yourskey) on yourtable as part of your jobstreams.

There is a utility called format.conv (or fnuxi - a synonym) that has -import and -export (and other) options for performing exactly this task.

Initial information can be obtained by executing the command HELP FORMAT.CONV from a dssh prompt on the server; full information exists only in UniVerse manuals, which can be downloaded from IBM.