DSXchange

Without knowing your job design it is impossible to say. Dumping the score may help, when you re-enable operator combination, to identify where a cycle may be being created.

That seems reasonable. There was no error in the stage per se.

I vaguely recall that there is a "treat warnings as fatal" check box available. Perhaps you could use that.

Oconv(TheNumber, "MX") will convert a decimal number into hexadecimal equivalent.

This has no relationship whatsoever with ASCII/EBCDIC encoding.

In server jobs the Aggregator stage has First and Last as set functions per group.

I always counsel never to design server jobs to abort. Issue warnings, certainly, but retain control.

So can anyone else. U is the name of one of our posters. The second person personal pronoun in English is spelled "you", not "u". Please strive for a professional standard of written English on DSXchange. It assists those whose first language is not English, if they don't have to...

Sort by customer_id, weight and last_update_dt and use Remove Duplicates stage to preserve the last of each customer_id group.

Add a column to the target table indicating the source table. Provide this value via a job parameter that can also be used in the table name field. The name of the table could be generated inside a loop in a sequence.

Import the target table definition. Are there any NChar or NVarChar columns defined in the imported table definitions? If so, these are the ustrings about which the exporter is complaining.

A simple cat command (Execute Command) upstream of the User Variables stage should suffice. The user variable derivations can use Field() functions to extract the relevant columns from the row that was read. (I've assumed that the file has one line. Without that assumption the requirement makes litt...

But the transformer in server does not alter the sequence of the incoming rows but the parallel job does. No it doesn't. YOU may have done something to affect the sort, such as incorrect partitioning or specifying sorting on the input link, but the Transformer stage does NOT affect sorting by itsel...

A useful hashed file is SDKSequences. You can access this through the SDK key management routines (among other methods).

Arnd, I'm afraid that haiku has 18 syllables.

Perhaps the final line could be "no longer green".

You need MQ client.

The MQ server can be local or remote (which is a change from version 7).

You never use wave number. It is for internal use only. It is DataStage's allocated number to identify a unique instance of a job run.