DSXchange

Hi Folks,

Can anyone tell me the difference between server job compilation and parallel job compilation.
i neeed more information on the server job compilation.
How the sequences were generated in server job in the earlier verisons when we didnt have surrogate key generation.

Thanks

There is no single difference between server job compilation and parallel job compilation, as your question seems to imply. Nor is it very clear just why you "need more information on the server job compilation" - is there any specific thing you have found in the manuals that is not clear? The way that job sequences are compiled is largely unchanged - whether or not surrogate key generation is involved is irrelevant to a job sequence, but is relevant to the actual jobs that use surrogate key generation.

Server job transformer stages compile to DataStage BASIC source code which is cpmpiled as a routine and called by the internal routine DSD.StageRun when the job runs. Everything else becomes records in the "job configuration" table which is used as data by the internal routine DSD.RUN when the job runs.

Parallel job transformer stages compile to C++ source code which is compiled and linked and called by the Transformer operator. Other "out of the box" stages are converted into Orchestrate operators (instances of particular classes in the Orchestrate framework) and a script indicating the job design is constructed (this is not the artifact that is run, however, as it includes no information about parallelism or the configuration file).

There's a lot more, but you only asked about compilation. And until you buy a premium membership you aren't even reading this far.

Can you please list 5 differences between them?

Can you please list 5 differences between them?

Once you can read the entire post above you will have the whole story.

Please note that neither I nor the other four premium posters is paid anything for posting on DSXchange, but I do want it to keep going, which is why I'm always advocating that people buy premium memberships - one of the ways that the site is funded.

Okay Ray i will register soon....!

Nagaraj wrote:Can you please list 5 differences between them?

If this is an interview question I am surprised that it is in the list, that is like asking "list 5 differences between compiling COBOL and Java". Any answer answer means nothing towards being able to use either compiler, or either Server / PX.

It is not an interview question.....i wanted to know about server jobs background...! thats all.....! i am doing projects on parallel jobs...i was thinking why cant i do some jobs in server, data volume is very less for us...!

You can do jobs on server, unless the politics of your client forbids it. There is no technical reason not to, and there are some technical reasons why you should, particularly for small jobs. The main one of these is the far smaller startup time of server jobs.

okay great thanks

Nagaraj - I'm sorry for assuming that it was an interview question; usually people wanting to know such information ask "Can you list some differences between them" rather than "List 5 differences"... I'm sure you see the difference in how it comes across.

Pretty natural assumption when asking for "5 differences" rather than "some differences" or "the difference"... very quiz/interview like.

Yes i can understand, thanks.

Hi Nagaraj,

Once we compile a server job, a basic output code will be generated.
Once we compile a parallel job , OSH shell cript will be generated..Also C++ code will be generated for some stages.

Thanks
Deepti

Thanks Depthi for the information.