DSXchange

Three questions, one answered.

How do you expect us to help us if you don't provide the requested information we need to do so?

For best value this FAQ would need to be linked to from the Signup page and from the welcome email.

Should we require an "I have read and agree to the terms & conditions" check box?

Its severity is warning not because it's a failure, but because it is an unusual event. It occurs only the first time one of the KeyMgt... routines from the SDK is used. You will never see this message again in this project. SDKSequences is a hashed file in which these routines preserve the next val...

But it still has to DO the sorting and aggregation, and the DataStage clock is still running once that request has been sent.

What value do you want to pass?

Is the first line in your source file column headings?

Try using a stage variable, perhaps called svFirstTime, that is initialized to @TRUE and reset to @FALSE once a row has been processed. You will need a second stage variable to remember the prior row's value.

Use two separate Transformer stages. It's easier to understand, and you don't need to muck about with expressions. The first traps your business rule rejects; the second traps database rejects.

As a general rule we ALWAYS choose integer for foreign keys in star schemas, because they join the most efficiently.

Probably not. But there IS TimestampFromDateTime(date,time) which is documented in the manual.

I tend to use Modify stage for data type conversions. In the Modify stage one uses timestamp_from_date[time](date)
(Or omit the time, in which case midnight is used.)

It may need two - DateToTimestamp(StringToDate(...))

I vaguely recall someone mentioinng at last year's conference (or maybe the previous year) that they (IBM) were working on a utility for this. Get your vendor to chase it up - maybe your vendor can even get on to the product manager (Keith Kohl).

If you are planning to use CRC32 to detect change, don't. Parallel jobs have three different change detection stages - the one you choose depends on what format you require their output to be. The stages are Difference, Compare and Change Capture. None of these risks the false positive that CRC32 ca...

That, too.

lstsaur wrote:Two weeks to solve this kind of problem? Unbelievable.

Unbelievably fast, or unbelievably slow?
Bear in mind that the support folks are working on other people's queries too.

Or two lookups, against the same Lookup File Set. You will need to finagle the key column names, but that's no big deal - you can do it with Copy stage.