DSXchange

In addition to what Arnd advised, your method may need to handle changes in time within a time zone (usually called "summer time" or "daylight saving time"). Sometimes the change dates are regulated, some places it's more erratic. And some countries involve more than one time zone.

Welcome aboard. Is this a server job (as marked) or a parallel job (as posted)? The answer will be different depending upon which it is.

Why do you need a loop? Sort the files into date order (for example using an ls -lrt command) and use that result as the command line for a cat command. For example cat `ls -lrt /path/filepattern` Use that as the Filter command in a Sequential File stage. If the file names don't all fit on one line,...

Accumulate the non-key columns into a value mark (@VM) delimited string and write to a hashed file where that is keyed by the grouping key and that column is defined as multi-valued in the metadata. Then read from the hashed file normalizing on the multi-valued field.

On a pedantic note, VarChar(1) is a waste of space. Char(1) should always be preferred, otherwise you're carrying the length prefix that has a constant value.

What has changed? "Nothing" is not the correct answer. The change may be in DataStage, may be in Oracle, may be in the operating system, may be anywhere. You must become a detective.

Why?

Would you use a hammer to fasten a nut and bolt?

Create a job that SELECTs the current value then passes this to the Oracle table using a regular INSERT (one with a VALUES clause containing a parameter marker).

DataStage wants to pass a value for at least one column for an INSERT.

Are you prepared to use a Server job? It's easy there.

Have you checked the PXPivot stage? (as opposed to the Pivot stage)

Try here first.

There is a company that exhibited at IOD EMEA 2010 that has a utility for converting BTEQ scripts (and lots of other things, like Informatica, Ab Initio, etc.) into DataStage jobs and metadata that can be consumed by Metadata Workbench. The name of the company is COMPACT (there are more details, but...

Hashed File Calculator is still supplied (as at version 8.1.2).

Exactly the same problem, on identical hardware, using identical databases, identical job design and identical software versions?

If not (and note that you have no knowledge of the original poster's environment), please begin a new thread.

The local manager of VMARK (first vendor of DataStage) here was fond of overclocking the CPU, much to the annoyance of the IT people who had to replace his fried CPU from time to time, but it did go faster! Systematically investigate your design to determine whether you're doing anything you don't n...