DSXchange

Messy. To get a commit to happen on every row just set the transaction and array sizes to 1. They are optional options on most parallel DB stages and default at 1000. That way you can be sure the row is in the database before you query it. I assume you are doing this to retrieve a generated sequence...

I would say no interaction possible between DataStage and Rochade however good interaction between DataStage-MetaStage and MetaStage-Rochade. DataStage Hawk has more metadata processes built in and under that release you may have DataStage-Rochade capabilities. DataStage Hawk is also a more open rep...

Agree, DataStage server and enterprise jobs using the XML Input stage do not handle large XML files very well. DataStage TX is more effective with XML. You will find performance is much much faster if you can process it as a text file rather then XML and you wont hit a memory limit when the stage tr...

Probably your after-job routine. Reset the job and have a look for a log message starting with "From previous run..". An abort caused by a BASIC code error is sometimes retrieved and displayed by a reset. Also try the Test button from the routine editor to view output messages from the routine.

But then he'd have to upgrade from server edition to enterprise edition because it's going to be a parallel stage.

You have posted in the DataStage TX forum but you are referring to parallel jobs, which belongs in the DataStage Enterprise forum. You can join two DB2 tables in the source database stage via a user-defined SQL statement, you can join them in a job via the join, merge or lookup stages. Refer to the ...

On a new project I log into Administrator as dsadm, go into project properties and check the "Enable Runtime Column Propagation for Parallel Jobs" and click OK. When I go back into project properties the option has reverted to unchecked. I am unable to turn it on for this project. Anyone know why? A...

Use a complex flat file stage to flatten the data. Switch between ascii and binary to see which one works using "View Data" to test it out. Think of this text file as a hierarchy of database tables and work out which records you want to retrieve from it. You may need to process it via more then one ...

I am on a project where we moved from environment variables ($ENV) across to project variables ($PROJDEF). The good thing about environment variables is that you can share them with other applications. For example cognos cube builds or ftp scripts. However we had a dev environment that was also used...

Easy to do in a custom routine or in a shell script. In a BASIC routine you open the file, read the first line, if there is no data to be found then you write out your "DO DATA TODAY" and close it. Call it from a sequence job via a routine stage that follows the job activity stages.

I don't like seeing data butchered by dodgy sql functions in an ETL tool. This reformatting should be done in a transformer rather then the SQL select to reduce the load on the source database, make the modification transparent to metadata reporting and give you increased control over rejections or ...

I did come across a few third party products that help in this area. One is the Unix to Windows bridge, it creates a bridge between a Unix server and a Windows server and gives the Unix operating system access to all the ODBC drivers on that Windows server. This will in turn give you access to any d...

I don't know why you would put the surrogate key generator into a shared container. A single stage within a container does not make much sense. You pass job parameters into a container by ensuring the container has a job parameter that receives the high value from the calling job, which also receive...

Use char(10) in a transformer. Eg. input.col1 : char(10) : input.col2

You can build your own parallel processing capability using a multiple instance server job. The tricky part is partitioning the data as it is a sequential source. Option 1 - chop up the file using operating system commands. Option 2 - partition the data within the job using a transformer constraint....