DSXchange

This requires a classic fork-join design.

The DataStage TX product was moved out of DataStage some years back and is now called WebSphere Transformation Extender. Your question does not relate to that product and consequently does not belong in this forum.

<relocated>

Why are you posting this question in this forum, which is about the DSXchange site itself and which, by the way, is not backed by an Oracle database? We need to know the appropriate job type (server or parallel) to be able to offer the correct answer. This is achieved by posting on the correct foru...

The apparent format of dates, times and timestamps is determined by default via settings in the Administrator client. These can be overridden in jobs and even in functions such as StringToDate().

Look at the "bad" data. {0d} is a carriage return. I'd guess that you are reading a file that has DOS-style line terminators but have specified UNIX-style line terminators in the format properties of your Sequential File stage.

Check for lock contention in any database to which the job connects. A job will wait up to forever for a database lock (or table or row lock) to be released.

ArndW wrote:I gather that this behaviour was different at releases prior to 7.

Yep. Heaps slower, for one thing.

Not an anomaly - a deliberate design decision. Storage in Data Sets is fixed width (the same model as COBOL) so that addition and retrieval is in constant-sized chunks, which is very very efficient (the ZMOV instruction or equivalent) rather than needing to calculate offsets.

A bit obscure to introduce the SH verb, Arnd. A call to DSExecute can go directly to the operating system shell.

The initial value can be any valid value for the data type. 0 suggests itself in this particular incarnation.

Welcome aboard. The short answer is that you must handle them; you can not provide a null value to most Transformer stage functions. Your options are basically twofold; take an alternate action if the incoming value is null (probably using an If..Then..Else construct) or replace the null with some p...

If your data type is date you will need type conversion functions DateToString() and StringToDate() with different format strings. If your data type is string you will need Field() functions to unpack the components and Trim() to remove the leading zeroes, then concatenation to re-assemble the date....

What's the data type of drawtime? Assuming it's a string, you can use Field() functions, arithmetic (add 12) and concatenation to build the desired output string.

There is only one pivot stage in version 7.5.

Search on your client machine for a document called Pivot.pdf - this is the manual.

What you seek to do is called a "horizontal pivot" and is performed by the Pivot stage in DataStage.