DSXchange

You may need to apply conversion functions like StringToDecimal() to prevent warnings.

Write out a specification in English, since not everyone reads PL/SQL code. This may also help you to determine what you want to achieve, by forcing you to focus your thoughts. Think about testing your algorithm; test values, expected and actual results.

This is a server job.

I would not bother with stage variables at all. I would pass in the maximum value as a job parameter and use the expression

No, but it will in the next version (probably 8.5).

The code could alternately be invoked through a Routine activity in the sequence. Return all values in a delimited string - ideally a dynamic array because this can manage nulls - and, in the User Variables activity (or anywhere else) use the Field() function to retrieve a particular value.

Please mark thread as Resolved

When you mark a topic as resolved, our expectation (to help future searchers) is that you make a post explaining how it was resolved.

To pre-empt your next question, there is no alternative than to reinstall.

How large is the log? Is the log corrupted?

Please advise what the sequence does. Everything it does.

The second thing that comes to mind is that a failed lookup returns nulls from the reference input.

My guess is no. But you could write your own custom stage.

Try using an explicit Field() function rather than the delimited substring extraction syntax.

Of course it takes a long time - you have a correlated subquery here.

The only way you will get it to run faster in DataStage is to have both query results available - then DataStage can do a lookup on the unique instances.

Use a Pivot stage. Upstream of the Pivot stage use a Column Generator stage to add the constants. con_id con_sus_id con_id_K con_sus_id_K 1 10 T C 1 8 T C 2 4 T C 3 6 T C 5 7 T C 9 8 T C Pivot on con_id,con_sus_id and on con_id_K,con_sus_id_K

In a word: PATH