regaring universe

vijaykumar · Post by **vijaykumar** » Thu Dec 07, 2006 1:17 pm

Hi Gurus,

As we are into better understanding of universe, you people have helped me a lot, but still need to identify some objects in universe.
j\1\root
j\2\root
j\3\root
...
.....
....
iam getting LINE NUMBER 15 LINE NUMBER 16
LINE NUMBER 17
is not available for all the records.
in DS_ROOT

again it displays j\19\root.

what does the line number not available indicates.
why is that showing.

plz help me in better understanding universe.

cheers;
vijaykumar

vijaykumar · Post by **vijaykumar** » Thu Dec 07, 2006 2:34 pm

Hi Gurus,

probably i feel that was not enough information i have given to expect some thing.

what iam doing is that iam trying to extract data from universe, since my work extensively involves working with universe and i cannot find in manuals, even tried different other ways.

As iam extracting the data i identified some objects.

REFERENCE RECORD ID LINE NUMBER
J\1\ROOT OO1

C\1\ROOT 002
.. 003
004

after some records , iam getting reference record id for this number does not exist, why am i not able to identify reference record id for certain line numbers. As we are using server version on windows platform, can i able to identify these objects using parallel version. plz help me.

cheers;
vijay

...
...

ray.wurlod · Post by **ray.wurlod** » Thu Dec 07, 2006 7:43 pm

One of the reasons that metadata are not published for repository tables like DS_JOBOBJECTS is that the record structure is not the same for all records.

For example the reason that you can't find 018 in the ROOT record may be that field 18 is empty and you've chosen to use LIST.ITEM to display the record contents. LIST.ITEM skips empty fields. Or it may be that the ROOT record contains fewer than 18 fields.

And, it's NOT UniVerse. Hasn't been since version 5.2. It's the DataStage Repository.

What IS "your work"? Nothing in DataStage should require this level of detail. Users are *supposed* to work exclusively with the clients, APIs and command line interface.

vijaykumar · Post by **vijaykumar** » Fri Dec 08, 2006 3:37 pm

Hi gurus,
our company is doing the reverse engineering.
dealing with universe- which has developed jobs ,routines, procedures and so on... my initial step is to analyze the objects in the universe.
As no documentation is present to give the clear view.
you people are gurus helping out the developers with any sort of problem.

reference id linenumbers datacontained in linenumber
j/1/20.j/10/20, 001 NRO
j/11/20

j/27/20,j/28/20 002 CCONTAINER VIEW.
.... .....
.....

linenumber does 017

j/32/20 018.

what does line number here means.
As Ray said, i was able to understand, but can u explain what is meant by linenumber here.what is meant by data contained in linenumbers.

only you people can analyze these , as you are keen with every environment.Plz help me.

cheers;
vijay
not exists........

kduke · Post by **kduke** » Fri Dec 08, 2006 10:26 pm

Post removed per Ray's post below.

ray.wurlod · Post by **ray.wurlod** » Sat Dec 09, 2006 4:05 am

Reverse engineering is explicitly forbidden in your DataStage licence agreement.

Given that you've admitted that this is what you/your company is doing, I feel constrained from assisiting further in this thread. IBM's copyright police are famed for their vigilance and preparedness to litigate.

Knowledge is one thing. Crime is another.

kduke · Post by **kduke** » Sat Dec 09, 2006 7:56 am

Reverse engineering implies replacing this product with another product. Is that what you intend to do?

We use these methods to only better understand our ETL designs. We actually like DataStage and just want to be better DataStage developers.

chulett · Post by **chulett** » Sat Dec 09, 2006 8:57 am

I for one have always wondered what the motivation was behind all of these questions and have purposefully restrained from chiming in at all before this point. I mean, come on... this was more than the small amount of dog paddling some people do in the shallow end of the UniVerse pool looking for legitimate knowledge like 'How do I make a Read Only job readable again?' - this was full fledged floundering down in the deep end, seeking forbidden fruit of the UniVerse Dark Arts. And you could tell from the questions that they had just been thrown in there by their company, without so much as water wings! [splash] [glub,glub]

And Ray's "Why in the world do you need to know this level of detail" question has never been answered fully, other than the 'reverse engineering' comment recently made. This line of questioning, to me, seems to go way beyond any effort to 'better understand ETL designs' and has for quite some time.

I know we're here to help and hate to see people floundering, but...

vijaykumar · Post by **vijaykumar** » Sat Dec 09, 2006 4:07 pm

hi

sorry i used the different word to describe what we are doing.
what we are doing is.

source- universe.
target-oracle.

doing the normal ETl process.

i wanted to analyze the data in universe but i used the different word.

i never thought that source can be universe too so used that word.

plz assist me. iam doing the normal etl.

you are the gurus of datastage without you people assistance cant go further
just want to analyze the data in universe.

plz help me.unfortunately mispelled different word to explain my process

chulett · Post by **chulett** » Sat Dec 09, 2006 5:36 pm

No, I think you used the correct word for what you are doing and the questions you are asking. None of this is 'normal ETL' except in a very loose sense. I've been using the product since version 3.1 and have never needed to ask the questions you have been asking.

And your source isn't just 'universe' - you are sourcing the DataStage Engine Repository. Those structures are purposely not documented as 99.99% of the people on the planet using the tool do not have any need to be in there goofing around. Ray and Kim and a handful of others are... unique, let's say, having worked for the company that engineered the product DataStage is based on - the 'dot oh' release. That's primarily why they know this stuff.

Tell us WHY you need to 'analyze the data in universe'. What is your quest, your goal with all this? Once you get all of 'this' in Oracle - then what?

ray.wurlod · Post by **ray.wurlod** » Sat Dec 09, 2006 11:23 pm

You can use the Documentation Tool (Reporting Assistant) to dump most of the design-time information from server jobs into an Oracle database. They even provide Oracle scripts for creating the said database. Look in the Documentation Tool folder where your DataStage client is installed. Why re-invent the wheel?

I believe this will totally meet your re-stated requirement for dumping the design information of server jobs, without requiring any knowledge of the underlying storage mechanism in the DataStage Repository.

vijaykumar · Post by **vijaykumar** » Mon Dec 11, 2006 2:28 pm

Hi gurus;
you are always on the top, i feel that a datastage developer in the entire world should come up to the forum inorder to get precise answers.
i cannot think without this forum.
as ray suggested a reporting assistant.
can i dump all the information of jobs, metadata into oracle.when i dump into oracle , will it be in the format which the oracle database requires.
bz universe has even the all the built in routines, sequencers which were used in the jobs. when i dump from universe into oracle, will i able to know which routine was involved in particular job or the dependencies of 1 or more jobs.plz help me.

can u plz explain me

cheers;
vijay

ray.wurlod · Post by **ray.wurlod** » Mon Dec 11, 2006 3:30 pm

Look where I guided you. There are scripts there to create the requisite tables in Oracle.

us1aslam1us · Post by **us1aslam1us** » Mon Dec 11, 2006 4:15 pm

Further post your coments in the server forum.

sam

vijaykumar · Post by **vijaykumar** » Mon Dec 11, 2006 4:20 pm

Hi,
what i have analyzed is that by using reporting assistant, we can generate the reports from a developed job or project which is stored in repository.
but here the design is different.
i want to generate report from dsengine which looks like as much as universe.
iam following your advise and want to use reporting assistant as i feel this is the best strategy.
now the problem is how should i connect to universe and generate reports.
plz advise me if iam going wrong and help me in connecting with universe which has already bultin jobs,routines and sequences.

cheers;
vijaykumar