DSXchange

I don't see how that would happen. Post the details of your job design including constraints, stage variables, etc. As much detail as you can stand to document.

Well, there you go! Please mark the topic as Resolved using the button on the top of the page.

What's your target? Update action?

Put your source keys into a work table if not already there. Join to existing keys and build a hashed file. Reference lookup drives update v. insert and gives the 'old' value to append to for the updates.

Pretty standard stuff.

Change the Array Size to 1 temporarily and see what record is logged as the problem child. Array Sizes over 1 can confuse the logging process.

chulett wrote:When you use TO_DATE() all that matters is that your mask/picture matches the data.

Yours still doesn't.

eostic wrote:A good point lstsaur makes.

Ah... Ernie Ostic, the Yoda of DSXchange.

So, what approach are you planning on taking? And you need to answer Ray's question.

Odd, every post you've made so far has had the same subject - "datastage designer". You really need to put a teeny bit more effort into that please, make them germane to the topic being discussed.

Between DRS and OCI? The DRS stage set to 'Oracle' uses OCI so I can't imagine there's really any difference performance-wise between the two.

In other words, you can't. But you can export everything and then hack it up into individual pieces.

Nope.

You can name jobs anything you like - as long as they start with a letter and contain only alphanumerics or underscores.

What's your exact version of DataStage?

It's more of a third-party ODBC driver thing, you need to help DataStage find them. You would have had the same issue with HP-UX.