DSXchange

the "ps" information for that process also shows the parent process id, grep for that id and you will (eventually) find a process with a recognizable job name or job number, which can be used to find the job name.

When dealing with NLS you always need to know exactly what encoding your data is in. In this case you need to find out what your source DB representation is.

None of the datasets you've mentioned for Teradata support Japanese characters so it will not be possible to store these characters as text.

You cannot create a new column and assign a default value in a modify stage, such columns need a value derived from some existing column(s). Is this what the OP meant in the original question?

Instead of explaining the compile/load process and as I can't get to a windows cpp compile to test the math library, let us make it simple. If you replace
with
does the error remain the same?

What is your database, is it DB2? Are you calling this using DB2command? Perhaps it would be wise to consult your DBA regarding how to execute the command and whether or not it needs to be executed.

This is a one-way route, one needs to reach the OSH script in order to get an idea of what the job looks like. The OSH scripts pre-date the graphical front end.

Does the link back to the start loop activity really reference the correct start loop activity? If so, try deleting the link and re-creating it.

Have you tried to determine which user this might be? You could use Director -> Cleanup Resources to track down the record lock.

Could you give us a hint - does the command give a syntax error? I would suggest opening up a new thread on the topic, as this one is not only over a year old but already marked as "resolved".

Things were much simpler at Version 7 when this type of problem occurred, one could check the UV.ACCOUNT hashed file, ensure that the schema was deleted and then merrily get rid of the UNIX directory and re-create the project. At Version 8 and above the original entries exist but one needs to deal w...

The length of a CHAR( n ) field is always n , regardless of how many spaces it contains and get TRIMmed out. What the transform c++ code might be doing is an implicit conversion to VarChar() of a Trim(CharacterString) function result and then doing a length computation on it. It might be that differ...

Is that the only start loop in the job sequence?

The error message is pretty clear, your end loop activity doesn't reference a corresponding start loop activity. Do you have a link going from the end loop to the start loop?

I don't quite understand your post - could you please explain how your modify stage uses "pdct_num"?

What is wrong with trying my suggestion in my last post?