DSXchange

No. While reading the sample file directly can be "ok" you really need to get your hands on an ".xsd" which is the "schema definition" file. Any chance of requsting that from whomever is creating the file?

This is the error you need to correct: ##E TCOS 000003 18:35:36(001) <main_program> Syntax error: Error in "transform" operator: Error in operator arg: In field "ERROR_RECYCLE_COUNT": Parsing parameters "0,10" for schema type "decimal": Precision must be > 0 a...

All Ray asked was "OK, what has changed?" Obviously something has and unfortunately you are the one that's going to need to chase down the culprit.

It can't open the input file. Message:The primary document entity could not be opened. Id=C:\IBM\InformationServer\Server\Projects\EDW///Svkcdisd01/Projects/EDW/Outputs/sample.xml I'm assuming that's not the full path to the file you want to parse. Any idea how the "EDW" path is getting ad...

First question, in my mind, would be to confirm how your XPath Expressions were generated. Automatically from an xsd?

There really isn't the concept of "select *" in Server since everything is metadata driven. So the typical answer would be one job per table.

How many tables are "multiple"?

Apparently needed / created / used by the stage and (I assume) the user running the job didn't have permissions to delete it, so you needed to handle it manually.

I prefer them as well, for whatever that is worth, but tend to use a mixture of all three depending on the requirements of the job. But primarily - sequential files.

For that you'd need a level of job control (like a Sequence job) that could pass those values to existing parameters. Otherwise you will (AFAIK) need to change the jobs to include the Project parameters, set them to $PROJDEF and then update the Oracle stages. Pretty cookbook change and not that larg...

You should be using dsdesign.exe not dsjob.exe and you need to double-check that you are passing it enough parameters. Search the forums here for "dsdesign" to see examples of how it is called.

As noted, use standard SQL statements to join tables just like you would in any other database.

Or that.

etlbets wrote:If I could join DS_AUDIT with another table in Universe that would work as well.

Well, that is certainly possible. Join tables there just like you would anywhere else, use standard SQL syntax rather than the UniVerse Retrieve syntax that I assume that "sort" is part of.

Also helps a great deal if you have an xsd for the xml file and have imported the metadata from it, rather than from the xml directly or typing the XPath Expressions by hand. FYI.