DSXchange

Hello "ysr" & welcome to DSXChange! There are a number of ways that you can transfer a job log to a flat file. The simplest method is to use the unix command line dsjob -logsum with appropriate sub-parameters. This outputs to standard output and you can redirect that to a file. If you use the se...

Kumar, you can also look at the CASE statement. This is somewhat different in execution than other CASE statements (C, Pascal, Ada, etc.) in that the condition is queried on each case and can involve different variables. I've found that a well-structured CASE is easier to read than a complex and dee...

The file size limits depend upon your OS (AIX in your case), the version and perhaps the file system type. You will need to read the UNIX documentation for your version to see what the file size limits are, but if you are using a recent DS version I will assume that your system can handle files >2Gb...

Doma,

this might be the equivalent of the rollback segment filling up. What is your commit frequency for this job? If you check the generated sequential data file(s) are they at a logical limitation size such as 2Gb (perhaps the DB/2 read program uses only a small pointer).

Kumar,

I'm housing with 2 Brits who are avid sport fans, and I now understand a bit more terminology so I actually know what you meant. Don't know how many overs I'll last, though Never catch up to Ray, but he prefers baseball anyway.

jlock23, I found this reference in the book, but it was referring specifically to the dependencies attributes. I think you might be mistaken about being able to use this, perhaps you could say which PDF document on which page you found it. (if it's Server Job Developer's Gui de on 17-11 you found wh...

I can't recall what the error message is when your UNIX client doesn't have the correct environment settings, but if DS is up and running then that might be your problem, run the dsenv script.

Naveen,

2 steps.

1. In your job declare a parameter called FileDate
2. Create a Sequence, in that sequence define a Job Activity to call the job in (1). In the parameter derivation for that call you include the OCONV() functions that compute the name at runtime.

Naveen,

then you need to add a OCONV(@DATE,'D4-YMD[4,2,2]') or the like in the Sequencer.

Pass in the date part of the filename as a parameter to the job, and in the calling sequence use the EREPLACE(OCONV(@TIME,'MTS'),':','') functions to give the parameter it's value.

How about "Alcohol and calculus don't mix. Never drink and derive." or "I don't have a solution... but I do admire the problem."

Kumar,

in that case write a sequencer with just your routine in a RoutineActivity stage.

Trying to reverse-engineer a packed field is no fun

BTW, I just noticed that I broke the 10K barrier. Now I might rate a tag line as well. I'll have to dig out my Goethe or Nietzsche or perhaps Heidegger and come up with an excellent line in German.

Craig - who's pack ing I intentionally avoided that minefield, I though that Kumar has enough problems without my adding to them. I once got a copybook with several hundred columns that matched the actual data file .... except all the columns were shifted around ... so I can sympathize with the post...

The easiest way would be to create a server job that contains only job control (BASIC) code. This way you could use all of the DataStage job functionality through the UNIX/Windows command line call of "dsjob". The other methods require a bit more knowledge of UniVerse and are more effort than that l...