DSXchange

One possible improvement is identifying any single job bottlenecks in the process, a time when a small number of jobs are processing a hugh volume and the CPUs are not being fully utilised. You can fix this by using a multiple instance job and puting in your own data partitioning. Several instances ...

Duplicate thread, I just made the same suggestion as Craig a few seconds ago! How did you get "9369 WTC-CIF Real Asset"? This looks like a concatenation of values with a - added, do you have some type of output stylesheet applied? It seems odd it would concatenate values with - as a seperator.

One trick you can try with XML output is to create a sample XML output file in notepad, similar to the one you created in your example, then import it using import XML table. The importer will convert it to a set of column definitions which you can load into your XML output stage or use to set up yo...

There is a FAQ forum topic on removing duplicates.

Your local Ascential support office can send you the latest version. It may have been updated since I uploaded it to devnet since it was a work in progress.

If you can execute DTS packages from the command line through some type of db connection command then you can run a script from a DataStage routine via the DSExecute command. If you can run packages from a SQL Server stored procedure you could try using the ODBC stage to execute the procedure from w...

First off, move it from Windows to Linux or Unix. Now as to better performance it depends entirely on what you are trying to do, how many rows, source and target stage types, number of nodes, location of the node scratch and resource drives, partitioning of source and target databases, efficiency of...

From memory the row merger stage doesn't let you concatenate to two target fields. It will try to concatenate to the one field so you can't really do your codes and your descriptions into seperate target fields, unless you concatenate all 10 together and split them in a transformer.

That's right. Either recompiling or resetting your sequence job will clear out all checkpoints. You will see the status is aborted/restartable which means you can click on the start button without resetting.

Trainee9999, remember this is beta software, you really should be reporting it to IBM support first via the beta bug report email address in your release notes.

During the installation you get an installation options window of what components you are loading: Client, Engine, Domain, Repository Store, Documentation Have you asked support whether you can rerun just the Engine and Domain after an installation failure? I would expect the Documentation and Repos...

I did a dsxchange search and discovered I had used the "I have heard that" phrase at least six times so now I feel I must defend it. I'm of the opposite opinion, I think forums are a great place to talk about rumours. Why don't we have a rumour category? Better to let people debunk or confirm them t...

Anything under 10,000 lookup reference rows is quite small and will fit into memory so no need to go to the effort of preloading lookup filesets or datasets, just turn your 1000 row sources into straight lookup sources. Lookup stage should be faster then the join or merge options. Very tricky to mod...

It's not a parallel stage, it's a BASIC stage trespassing in a parallel job, so I guess it doesn't have the data partitioning required to run across multiple nodes. You probably want to run these jobs on a single node anyway as the BASIC transformer will be a bottleneck. The answer is to rewrite it ...

Maybe you should get a price on a WebSphere Information Integrator license, this will let you write SQL to compare the two tables across database platforms. It also gives you a lot more options for your DataStage jobs such as unstructured data sources and cross database query sources.