DSXchange

Yes, he is referring to that section. In addition to having a different shared memory segment base address, all of the clients that connect to DataStage need to differentiate between the instances by using different port numbers, which are appended to the connection ip address.

The invocation ID is a string value, and you would be retrieving from a sequencer, so you might have timing issues there as well. You can use a combination of system date/time and the PID of the sequencer. I'm not at a DS machine now, but I think you could get this either through a command stage or ...

Abha,

the cobol copybook merely defines the structure of your file. You need to retrieve and manipulate data to write it to this file. Usually you will use your metadata from the copybook in a sequential file write stage and add data to the file.

Abha, as Manoj has already pointed out, the code portion you added to your post is just the initial header portion of a job. It sets the values to some constant and variables and attaches, or opens, a job for use. The programming language is called BASIC and the complete manual is on your DataStage ...

I didn't realize that this was a PX post - sorry about that. As Ray stated, a hex value would work as this is just a binary representation of the internal format. You could also declare it as a CHAR(3) field and then take the value (Seq(In.Col[1,1])*(2**16))+(Seq(In.C...

pstefani, I am not really sure what you are trying to ask here, but if you were to write your multi-instance job that reads your table and pass in a parameter which is used in your SELECT clause, in one job it would be "where C2 = TRUE" and the other "where C2 = FALSE". Start this job in parallel fr...

the UNIX fork() splits a process. An unsucessful fork() can be caused by various resource-based problems - the most likely one when working with PX is the maximum number of concurrent processes per userid. This is most likely the cause of your (sporadic) problem - each flavor of UNIX works a bit dif...

If your data is unsigned, then using the sdk function DataTypePicCompUnsigned ought to do the trick.

What is wrong with the generated code?

If you take a copy of the generated SQL from the SQL builder and execute it in your tool of choice (sqlplus, Toad, etc.) does it execute correctly?

We can't answer that, as there are too many factors. But why don't you try it in your system and get a definitive answer? Oracle -> Oracle. Hmmm, if they are in the same instance/schema then it might be fastest to stay within Oracle and do the copy. You can make the job multi-instance and add a para...

Even though loveojha2 has answered your question, it wasn't a complete answer. The log files for DataStage jobs are stored in hashed files, which are not accessible from UNIX. There are different ways of reading these files or dumping them to text files (the dsjob unix program lets you do this with ...

In addition to the target, what type is your Source?

koolnitz wrote:... What say?

Can't say anything more except what I suggested first - ask your DBA to monitor the table while you run it to find out what is happening. The error itself comes from DB/2 and needs to be analyzed there.

Most likely you are getting a count() returned value of 0, which the Field() function doesn't like. If you reset the job you should be able to see your exact error in the "from previous run" log entry.