DSXchange

You will connect to MS-ACCESS using ODBC. If your DataStage server is also on Windows then you can use the drivers supplied by Microsoft and use MS-ACCESS directly. If your DataStage server is on another platform you will have to find a source for ODBC connectivity, as those drivers are not part of ...

Did you actually code a reject link in your output stage to see if some error is being found on the execution of the statement?

Normally I wouldn't see a problem with user-defined SQL going to a different table in the same instance.

Can you reduce your data volumes to just 1 row in each input file and post your data as well as dupilcated output? Also, could you try setting $APT_EXECUTION_MODE to sequential and use a 1-node config to see if the data duplication still persists?

Do you have a 10-node configuration and have set the partitioning on the transform stage to "Entire"?

If you do the lookups in a couple of transform stages instead of just one and set your job to use interprocess communications then you will see a marked performance improvement as each transform stage will be its own process. This won't help on a single CPU machine.

That is hard to tell, but I would think that passing the job the names of the sort columns should be enough and that you wouldn't have to add an extra temporary column to the data in order to sort.

Go to the generated code for the transform stage and see what datatypes and maths are used. Since floating point numbers cannot represent all values small errors can creep in, usually the use of a representation with more precision helps but can never remove all imprecision. In your case you know bo...

They are completely different entities. Environment variables are UNIX (and Windows) constructs that are used for user and OS information to be stored and used; i.e. for keeping information on what PATHs to use for programs and libraries and for storing program settings such as the ORACLE or DB2 ver...

Welcome to DSXchange. You will find that the people here are usually helpful, but the use of "urgent" in the question usually has the opposite effect since we are all volunteers here - it might be urgent for you but not for anyone else. Your use of urgent in the subject line makes things even worse....

Building a good solid generic job or shared container in PX isn't as easy as it would seem. But by using rcp and proper programming techniques it is a very powerful and useful thing to do.

The deletion cannot be done from the command line. You can check the status and run dates/times for a job in the director, and use usage analysis in the manager to see where a job is used.

Use the standard methods such as rcp, scp, ftp and nfs mounts

I would recomend using the UNIX split command and splitting the file into n-line or n-byte chunks. It is fast, effieicent and a builtin function in every UNIX flavor.

Do you have a SELECT clause in your hashed file stage where you are donig the 'view data'? It seems that keys that don't match the pattern 1X-3N are not being processed. Could this be happening at the time you are writing to your hashed file? If you write to a sequential file instead of the hashed f...

abc123 - run it with one node in TEST, not in production. This will most likely expose that your issue has to do with partitioning, since I assume that it will work correctly with just one node.