DSXchange

A trigger on the source table could invoke ETL processing.

A SortMerge collector is nothing to do with a Sort stage; it is the collection algorithm specified on the input link to a stage - probably a Sequential File stage - that is executing in sequential mode.

shirishareddy wrote:If we set the Condition Not Met=FAIL, job will fail.I guess we need to give the Condition Not Met=DROP.Correct me if i am wrong

You are wrong.

This is because the Condition Not Met rule is irrelevant to the original question.

You may need to look on scratchdisk to find the log, if you have not specified a location for the sqlldr log file to be written.

Three Lookup stages with the second and third being conditional lookups would be one way to solve this, otherwise one Lookup stage with three reference inputs and a Transformer stage downstream of that to test for the null columns.

ODBC.H does exist in UNIVERSE.INCLUDE SELECT @ID FROM UNIVERSE.INCLUDE WHERE @ID = 'ODBC.H'; So it's not that. Is this error reproducible? Perhaps you simply lost the connection from that client. The error is occurring when you are editing the routine, not when you are trying to access the database....

Sreenivasulu wrote:I think sooner or later datastage would need to become fully webbased

I have no idea why you think that but I, for one, would oppose any such suggetion vehemently. Most of the serious "heavy lifting" is best done in batch using very high speed server-to-server network connections.

If you initialize the stage (see my earlier post) it will initialize the state file (assuming that the stage has a state file associated with it).

It's not possible. The behaviour you have observed confirms that. You are specifying a 1MB buffer on every link in the job - that should be more than enough for most jobs. Buffers do not fill - they are continually being written to and read from. They simply allow more rows than one at a time to be ...

What are the transaction and array sizes in the Oracle stage that writes data into the database? If these are too large they can cause this problem.

The error message suggests that the last line in the file lacks its record terminator. I agree with Craig's comment about record terminator; you need to use the Record Terminator String property set to "DOS format".

You can initialize the stage by creating a job in which the stage has an input link.

DSR_EDIT is an internal helper subroutine called, among other things, when you are editing something in the Repository. Can you edit routines in other projects successfully? Try re-booting your DataStage client machine.

This is a "general" repository error, meaning that no other information is available. Can you:

• (a) run jobs from DataStage clients in this project?
  
  (b) run other jobs in this project using dsjob command?
  
  (c) run jobs in other projects?

Perhaps your teams have designed things so as to preserve/stage data in Data Sets. I would consult them before proceeding.