DSXchange

Hi Mike,

I think my new approach, to base the search on a dsx, will be more future proof. Expect the new versions to operate this way. So yes, I use the tools myself, and do plan to continue supporting the future release of DataStage.

Chuck

I think the easiest solution is to create a second job that appends a single row to your output file. This could be as simple as a transform outputting to a sequential file. Use a stage variable as your constraint, and allow it to be true only for one row. Example: Initialize the stage variable to 0...

Check my web site, www.anotheritco.com. Under DataStage Tools, I have some Universe Basic programs that may help. They work at release 5 and lower. I also have a release 6 and above routine that does similar things by searching a dsx, but it is not on the web site yet.

Chuck

Let's back up. Can you use a job parameter instead of a stage variable? If so, then including the parameter in a file or table name is no problem. Basically, enclose it in # within the file pathname (e.g. "/home/login/folder/#duedate#/xyz.txt). Now, how do we get the value passed into a job par...

Although the "execute command" stage does not like parameters, you can easily replace it with a "routine activity" stage, and hide your command within it. Use DSExecute to run your command from the routine.

Chuck

Just because you use an aggregator, does not mean you have to use a group by. That is one of the nice things about "Last". This is also a handy trick if you output based upon the difference between the previous row and the current row. Since the last row will never have something to trigge...

I suggest you combine the two approaches. Calculate your totals using stage variables, and feed the stage variables to an aggregator with the "Aggregate function" set to "Last". When the transform completes, the aggregator will output one row of your final totals.

Chuck