DSXchange

Josy, In the constraint, instead of marking the row as the reject row, you could do something like this: DSLink4.DBMSCODE <> 0 You can find DBMSCODE by right clicking in the constraint menu, pick Link Variables, then Outputs, click on the plus sign next to the link name and you'll see DBMSCODE. Good...

Srinivas,

I'm not Ray, but perhaps I can help.

I suspect you need to "source" the dsenv file before running 'uv'. From Unix prompt key:

cd home/dsadm/Ascential/DataStage/DSEngine
. ./dsenv
uv

That may help.

Tony

I just ran into this problem with DataStage v6.0.1. I contacted Ascential support about it. Eventually we worked out the exact scenario that you mentioned above. The fix was to put the "umask 002" command in the dsenv file, (we only had it in the start up script (the "rc" file))....

Viswanath One way to do this would be to split the job into two jobs. The first job contains all the stuff that happens before the JAVA program is to be run. The second job contains everything that should happen AFTER the JAVA program runs. In a Job Sequencer job, put a JobActivity stage to run the ...

Thanks, Roy!

Thanks for the workaround. We're getting ready to upgrade to v7 in the near future and I thought this might affect existing jobs. From what you've said, you only see this when creating new sequencer jobs. Is this correct?

Thanks,
Tony

Roy,

Did you get a case number for this problem? I don't find anything on the knowledge base about this problem.

Thanks,
Tony

Thanks, Ray. I originally created just a "regular" hash file with all the default options using the hash file stage. When I've done that in the past, it created a type 30 file. I dunno what happened, but it's fixed now. Unfortunately, this happened a couple of months ago and the job logs a...

I thought I would give everyone an update. I found that a hash file was the bottleneck in the job. Further investigation showed that the hash file was creating a file for each row that was supposed to be in the hash file (with the file name being the data for that row). I had originally defined this...

OK, sorry it took so long to get back to you, I'm working on this in my "spare" time. I did try the Unix sort (4 million rows). It took 7 minutes to do the sort. Then I fed the results into the aggregator. It still took 4.5 hours to run. The next thing I'm going to try is to do the aggrega...

Thanks, Craig. I'll try the Unix sort and see if it buys me anything.

If y'all see a different approach to accomplishing my requirement, please let me know.

Thanks again,
Tony

"...Having you tried writing your data out to flat files, sorting them at the Unix level and then loading back to your Aggregator jobs from there? Might be worth tinkering with, or at least worth running a benchmark on one of your four streams. Still curious, how is the 4:30 split between sort ...

You're right, Kenneth. I missed that one. The output data should be Part Unit Unit ID From To Quantity 1 1 3 3 instead of 1 1 10 3 I'll try to explain the ranges. Say an assembly contains Part1, Part2, and Part3 when a Unit is originally designed. By the time the first 10 units are made, a problem i...

I was just looking at the job and thought I'd post some numbers. There are about 50,000 rows with From/To ranges. When these are expanded to individual Units, there are 4,029,542 rows that need to be sorted, aggregated, and put back into ranges. I have split these into 4 streams hoping to speed thin...

Hi Craig. Thanks for your reply. Yeah, I know it's a lot to absorb, but I felt like I had to explain things before I asked my question, so I would get a meaningful answer. Looks like the sorting and aggregating takes about 4 hours 30 minutes out of the total of 4 hours and 44 minutes. And yes the so...