DSXchange

You are looking to convert a string number into a VarChar(2) year value. I don't think you need to use any date functionality here. How long is the source? Can you do a right(trim(In.StringValue),2)?

Is the UniData instance local to the DataStage server machine.

Which header - the MQ header(s) or your message? What is wrong with a null? What is the issue that you are seeing?

... (Oct 22, 2008 5:20:46 PM), Install_repositoryinstaller, com.ascential.acs.installer.asb.InstallRepositoryAction, msg1, DB2 Location: /app/IBM/db2/V9 (Oct 22, 2008 5:20:47 PM), Install_repositoryinstaller, com.ascential.acs.installer.asb.InstallRepositoryAction, err, ProductException: (error cod...

What is the "WHERE" clause in your query?

The Search facility returned 43 matches on error code "80011". Do you think that your problem has a different cause and solution than all of these threads?

Using "urgent" or "immediate" in these forums almost always has the opposite effect.

The command line "dsjob -run" command will let you start jobs in any project (and on any machine).

You cannot avoid these files when using a BASIC transform stage in PX jobs. The error you received is most likely due to your PX job aborting, and the actual BASIC transform stage process still running. Since it is running the pipe isn't deleted and that causes problems. Try "ps -ef | grep [your use...

When you use the director to schedule it will create the entry on the server. What userid did you specify it to run under?

Craig - I think we do need to guess, since siauchun84 won't tell what the error message is.

My guess: column 3, record 54243 has a "h" character, but the column is defined as integer.

Write a job sequence that starts both concurrently.

This sporadic error can happen is the actual BASIC process is still running after the job completes. If this does occur, kill the process and delete the /tmp/ade* named pipes.

No. The hashing algorithm determines how the data is stored. If you sort your input stream by the key column you use for the subsequent lookup then you might gain a miniscule amount of time since each lookup will have a higher chance of hitting the same record, which is probably still in memory and ...

Compare your parameters between "good" and "bad" runs - are they identical? If you change the job to run in 1 node does it always work? Is there anything special about the missing data - i.e. key ranges, field values, etc. that might pertain to your selection criteria?