DSXchange

Perhaps it was reintroduced. Regardless, that would be a question for your official support provider.

From what I recall, older 8.x releases had this issue but it has since been corrected. What release / hot fix are you running?

ravi7487 wrote:Thank you for clarifying, But what initial value should I put in for the previous value?

chulett wrote:Whatever non-null value will not match with anything that might come in on the first record.

What kind of trigger is between "Job sequence3" and "Parallel job1"? Your job failure has nothing to do with your Sequence, you need to check the appropriate log to see what is going on. And by "appropriate" I mean the log of the job that failed.

As noted, you are not "manually" initializing them, you are properly initializing them using the mechanism provided for exactly that purpose. Make sure you understand, however, that it doesn't need to be a hard-coded value there, you can use an expression or function or job parameter there...

It would seem your only other alternative is to write a custom operator.

So... in other words, exactly what's already ready been posted as being used. :wink: Me, personally, I would rely on the expression automatically assigning the proper values, there's no need to do it "manually". Then treat the result as the Boolean it is. IsDup: DSLink2.deptno = LastValue ...

You've got all the logic you need right here. :? One issue could be your incoming data is not sorted properly. Another consideration could be the number of nodes this job is running on, if it is running on more than one node you'd need make sure you hash partition properly on the value(s) you are us...

Whatever non-null value will not match with anything that might come in on the first record.

ravi7487 wrote:for the first current variable, what will be the previous variable?

Whatever you put in the Initial Value of the stage variable when you define it.

You need to get the time as well, that or TimeDate() which gets both. And 1 hour = 3600 seconds, which is what you would need to subtract from the internal time, then convert back.

Switch the order of the stage variables.

You did ask how to solve it, hence my response. Last time I had the same question, I asked IBM. Some details could have changed in the last couple of years but should be close.

One answer would be to create a view from the overly large SQL and then source from the view.

All the various approaches here have been discussed ad nauseam. You can find them by doing an exact search for ETLStats as that always gets at least a mention in the thread.