DSXchange

I found it, if you go to the release matrix and click on DataStage on Windows it shows up as a Category 1 release. Releases in Category 1 Category 1 is defined as the latest release available for general distribution. Products in this category are eligible for all levels of maintenance support agree...

In an interesting progression of release numbers Ascential seem to have released 7.5x2 in December of 04 which runs parallel jobs on Windows, followed by release 7.5.1 in February of 05 which does not run parallel jobs on Windows. So you have to specifically ask Ascential for "the one that runs para...

Ascential set up a team of developers to convert PeopleSoft EPM customers across from Informatica to DataStage when PeopleSoft changed partners on the embedded ETL engine. I have heard this team had an Informatica to DataStage converter that did up to 90% of the conversion work, with the remainder d...

Not really a filter but a string transformation, have a look at the FIELD command in the transform function list or one of the string commands in the Modify stage.

This thread needs to be moved to the Server forum.

You can use the FTP plugin, this lets you start processing rows of data as they arrive on your server. I prefer to use an FTP script and bring the file across as soon as it is ready, then process them on the DataStage server as per the processing batch schedule. This provides for a more secure FTP p...

Good point about partitioning. Volume: The other thing I take into consideration is whether joining in SQL provides access to a filter that you would normally apply in the job that significantly reduces the number of rows you are taking out of DB2. For example 4 million rows in Table A, filter on a ...

Had trouble understanding your question. Are you talking about lookup up the same table twice in the one lookup stage? Consider using a source stage leading to a copy stage with two outputs leading to the lookup stage. The source is read just once but passed into the lookup stage twice.

Okay, I get you. We run nextval on every single insert row, so every surrogate is generated by DB2 and none are generated by the ETL job. It does give us unique values each time as the nextval command is supposed to increment the sequence.

You have a option within the parallel sequential file stage to "Read First Rows" that limits the number of rows read from the file. You would need to use a Unix command to count the number of rows, subtract 1 and then pass this into the job as a job parameter, set the "Read First Rows" to this value...

Yes, that will do it. The XML input stage will "flatten" your XML data into a single record type which can easily be handled by a normal transformer.

Is File 2 a subset of File 1? IE. Can you process file 1 and perform a left outer join/lookup against file 2? You do not seem to be aggregating anything from File 1 so a full outer join seems like an overly complex way to merge them. Try reading in both files and funnelling them together, then send ...

I would be more concerned with your scratch disk space then /tmp. Do your nodes have enough disk space to save datasets?

One of my personal bugbears is transformation within user-defined SQL. Have a look at it in detail, find out if it renames columns or runs functions on columns. This type of manipulation occurs outside of your DataStage job (on the RDBMS) and can be hidden from some data lineage reporting. To turn a...

I would never develop on a single node if my server could handle two. A single CPU server should be able to handle two nodes. If I had two CPUs I would consider four nodes. The more nodes you use the more likely it is that you will spot partitioning errors during your unit testing. If you cannot do ...