DSXchange

I am working on an ETL tool recommendation for a potential client and would love for DataStage to be the tool of choice. Currently they are considering DataStage, Ab Initio, and Informatica. The source is an DB2 database residing on an OS/390 mainframe. Target is a DB2 database residing on a Unix or...

I am working on an ETL tool recommendation for a potential client and would love for DataStage to be the tool of choice. Currently they are considering DataStage, Ab Initio, and Informatica. The source is an DB2 database residing on an OS/390 mainframe. Target is a DB2 database residing on a Unix or...

This problem was reported to Ascential with no solution thus far. However, we went ahead tried dropping back to the Complex Flat File version 1 available on the Web (version 3 is released with 5.2R1) and it is now working successfully. We have reported our findings to Ascential.

- Doreen

Hello Massimiliano There are many ways to create global persistant variables. My preferred method is to use a hash file to store any persistent data. You could of course use a manager "Routine" to open and close a hash file but why not use the tool? I sucessfully managed to create a persistent varia...

Hi Kasia, If appropriate for your job, you can avoid the need to have the filename as a job parameter by using the invocation ID to construct the file name. Its available as a job macro DSJobInvocationID. Like the other job macros you can use it like a job parameter - e.g. myfile#DSJobInvocationID# ...

Or, you can just save your job N number of times. In the situation where I have many source files of the same structure and different data all targeting the same target table, I try to concatenate all of the files into a single file, then use the following trick if you have a source sequential file ...

David, you will have to actively setup the job to handle this type of buffering. It has tremendous performance benefits, except when a job relies on a row traversing from a passive stage to another passive stage before the next row begins its journey. This is highly important if you are writing to a...

First modify the job and in the job properties screen set the "Multiple Instances" check box to true and make sure the file name and path are set up as job parameters. If you are writing to five different tables then the table name will also need to be a parameter. To run the multiple instances from...

Thanks guys. Seems I was under a misconception about COMMON variables. Im glad to hear they are shared amongst stages. Mike raises another interesting issue with row buffering. As I understand it, up until version 4, a single row is processed through all active stages until it hits passive stages, b...

Be cautious using global variables if you are using any of the row buffering techniques introduced at DataStage version 5 or beyond. Since rows are processed in a more pipelined fashion (multiple rows may in various states of execution concurently) the value in the Common areas may not be exactly wh...

If you create a function using a named common as Tony says below, those variables within that common are available to transformers throughout your job design. Any call to a function that contains that common statement will allow you to reference and change the variable. Changing said variable means ...

I would like to run 5 times the same job to load 5 similar files (same structure, but different name) into a unique Oracle table. Ive just discover the 5.2 version, and the multiple job instances functionality. How to allocate automatically different invocation_id to my job? Perhaps should I do it i...

I believe you can use a "named common area". The syntax is something like (going from memory here, so you better check in the basic.pdf)... COMMON /MemoryAreaName/ MemoryVar1, MemoryVar2, ... etc, MemoryVarN I have used this structure in calling 2 separate routines and sharing data between them (pri...

Comman Variables are persistent thru the duration of a JOB not between JOBS in the Job control section, you can define the common stmt as COMMON /Orders/PO, Name, Address, Total PO, Name, Address & Total being your Job Comman Variables that can hold values. you can reference these common variables w...

Except that every active stage is a separate process. Im fairly sure you cannot use COMMON to share values from one stage to another. Something which is "global" is job parameters, but you cant change the value of these inside a job. Within a transform stage, you can use stage variables. To share th...