DSXchange

chulett wrote:There are weeks when you don't wear your DataStage hat? [/boggle] Say it ain't so!

I also have UniVerse and Red Brick hats. See my web site.

I am trying to do an input DataStage job like that: 1. ONLY If the table does not exist, create it. 2. If the table exists already, insert rows without clearing. I know DataStage allows developer to input create table DDL and drop tables, but if I put the create table DDL there, the create DLL will...

Once you've done all those calculations, add 42% as a safety margin. Document it as a safety margin. When challenged, mutter something about the "ultimate answer (to life, the universe, and everything)".

Of course you can do it with DataStage componentry! Use a job sequence. Job #1 performs a lightweight SELECT from the table. Job #1 may abort if the table does not exist. Use a failure trigger to run another job that creates the table (perhaps even inserting zero rows into it, any maybe issuing an a...

Not wearing my DataStage hat this week. Maybe someone else can contribute in the meantime.

... of course, I meant DWH Architects, not DH Architects!

I presume the numeric piece is width (number of character positions within the file), but what do "AN", "BI", "NUM" and "P" mean?

This is a highly unconventional representation.

Manually, OK, but from where?!!

Somewhere there exists documentation, almost certainly in COBOL FD format, of what the COBOL program works with.

Basically, SELECT privilege from the system table is required so that the Oracle Enterprise stage can verify that the partitioning/parallelism information is in accord with what exists in Oracle, and can check that the executing user has appropriate privileges (INSERT, UPDATE, etc.) on the target ta...

I would expect, given that it's mainframe data, that you can get the COBOL FD from them - and you (or someone else) would probably have imported the table definitions into DataStage Repository from it.

A good topic to research would be Software Development Life Cycle - more on Google, etc., than here. The general rule is that, by the time things are in production, they've been fully tested and do not change there. If changes are needed, those changes are wrought in a development environment where ...

DH Architects are, in my experience, governed by consultation with Business Analysts who prepare target-from-source mapping specifications which, along with row counts from source, allows the size of the target to be estimated. But there's no way I'm going to put a figure on it, which your original ...

Welcome aboard. :D Your question is in the "how long is a piece of string?" category. The only valid answer is something like "big enough, and growing over time". I have worked with star schemas as small as a few Megabytes and as large as a few Petabytes. It's all driven by the business needs of the...

It's NOT "the full details". You have not told us on what basis the rows are to be combined. It is as simple as the first with the first and append "B4121 1 221" to all of them?

Checking the Omit check box manages your authentication by passing the Windows SID with which you are logged in to the client to the other machine (the DataStage server). Therefore, one would assume, that user ID will be the owner of any job created in such a connection.

Try it. Let us know.