DSXchange

I'm sure I mentioned that! Fixed width sequential files CAN be read in parallel mode. Each node reads 1/N of the lines in the file. Because it's fixed width (and therefore the row size and file size are quickly able to be calculated), each process can be given a specific seek() request about where t...

Welcome aboard! :D

Were you to search for 81002 on the forum you would learn that it means "connection refused or RPC daemon not running". So these are the two things you need to check.

The RPC daemon runs as dsrpcd; you check for this using

Type in dsjob just by itself to get all the major options. For example to get a list of link names you use -llinks . To get link information (including row count) you use -linkinfo . dsjob -linkinfo project jobname stagename linkname You then must parse the output, which has several lines. Remember ...

Looks like you've picked the wrong stage type. Investigate the Pivot stage for this particular task.

You can use the Administrator client command window or a telnet session to the DataStage server before invoking dssh there while attached to a project directory. DS_CONNECT is invoked as a command. It takes the DSN as its command line argument. It prompts for user name and password. You then enter S...

SELECT the value into a text file, use an Execute Command activity to read the value from that file.

The generated file name has a four-digit sequencer. This suggest to me a possible range of 0000 through 9999; that is, ten thousand files per processing node ne c'est pas? Fixed width sequential files CAN be read in parallel mode. Each node reads 1/N of the lines in the file. Because it's fixed widt...

DSLogEvent() is intended to write an event into a different job's log.
To write an event into the current job you can use DSJ.ME as the handle, as Arnd suggested, or use DSLogInfo(), DSLogWarn(), DSLogFatal() or DSTransformError().

Welcome aboard! :D

Does the same thing occur when you execute the same SELECT statements using the diagnostic command DS_CONNECT?

Can you successfully perform the query from a different client on the DataStage server machine?

If any row is processed on a link marked as a reject link in a server job, a Warning event is logged. This is how DataStage works. There is no mechanism of which I am aware for suppressing that Warning.

Test Sainath's theory by invoking ExecSH as a before-job subroutine to run the ulimit -a command and observe its output in the job log.

If you attach a job in a routine it is good programming practice to detach that job before exiting the routine. However, the job must be attached for the DSLogEvent() function to work properly. If your routine is - or appears to be - hanging, add some diagnosting statements to invoke DSLogInfo() and...

ISO8859-1 is not really a Japanese characters map. For example, it is a single-byte character set, while the number of different Japanese characters is definitely more than 256. You need to determine how the characters are encoded at source, and use this map between the source and DataStage. For exa...

In particular the Program Files/Ascential/DataStage/Templates directory must be writeable. This is where your job templates are stored. Yes, on the client machine.

yep I need to do the grouping on all 68 columns ! Eeek !!! In my experience, you get very little change after about three levels of grouping, even in the huge numbers of rows encountered in India and China. Sixty-eight?!! I'd be scrutinising that requirement with a very intense scrute indeed!