DSXchange

1)can we create readly only routine in datastage?
2)what is staging area in datasatge etl architecture?
3)what is syn-sort in datastage ?
4)what is the use of job parameteres how can u use them?
5)what are staging ,environmental,system variables?
6)how can u use datastage routines?
7) what is trim b and trim f functions in datastage?
8)what are the general problems u face when run a job?
9)how can you run datastage job in back_ground i their any command?
10) how can connect oracle to teradata?
11) how can use macro variables with examples?
12) why we look-up as hashed file why not odbc or flat file?
13)

I hope these are not questions in a job interview.

1) I would use Version Control.
2) DataStage itself is a staging area including hash files and sequential files.
3) Syncsort is a third party tool that sorts flat files very fast. Cosort is a similar product. All of this use plugins.
4) Job parameters are generally used to stored concepts that may change between dev, test and production DataStage servers like DSNs, user, passwords and temporary file paths. Normally you need to use #JobParam# to retrieve the value of the parameter.
5) Stage variables are only valid in one stage. They are used to store long complex repeatable derivations. They are also used to store values from a previous record. They can be used to accumulate values or aggregate values.
Environment variables are another way to setup parameters. Some say they are easier to port to from dev, test and production. Not sure if I agree with that statement.
System variables are assigned by the DataStage engine like @DATE and @TIME. Only a few of these can be used or assigned by a user like @USER1.
6) Routines can be used to do complex derivations. There are also before and after routines which can be used to trigger events at the OS level like archive a file.
7) Trim has several variations to trim more than blanks. trimb is trim back. trimf is trim in front. This way you can trim leading blanks or trailing blanks or zeros or whatever.
Run out of disk space. Run too long.
9) All DataStage jobs run in the background.
10) Plugins or ODBC is the way to connect to any database. The OCI stages are the Oracle ones.
11) I have not used macros.
12) Hash files are very fast for lookups.

Sounds like a job interview to me.

Kim, You're hired!

Mike

Mike

Thanks. I know we really do not want to help people get jobs when they do not know these answers.

Someone sent me a real nice email so I answered this.

How do macros work by the way? (12)

In my mind it is okay not to know it all. Only Ray knows it all at least when it concerns DataStage. Has Ray asked a question? I know Craig has. Not sure Ken has either. My brain is too old to hold all these answers. I am too old to care too. Now I am happy if I can remember where to look. First place to look is here. Thanks Ray, Craig and Ken. Just kidding around.

Kim,

I, for one (among hundreds/thousands of others I'm sure), appreciate your willingness to share your hard-earned knowledge and experience on this forum.

A good job interviewer will know the difference between someone who is just regurgitating an answer and someone who has actually internalized the concepts.

My understanding of the DS Macros is that they are just convenient short-hand so that you don't need to remember the API call syntax (e.g. DSJobStartDate).

Thanks,
Mike

Those macros. Isn't there some kind of macro built around job templates?

Ah. I didn't think of that possible angle. I haven't played around with job templates.

Didn't Ascential make an attempt at some kind of "Intelligent Assistant" or "Wizard" for jump-starting a job design? I guess that must be what you were thinking of in terms of a "macro".

Mike

Mike wrote:Didn't Ascential make an attempt at some kind of "Intelligent Assistant" or "Wizard" for jump-starting a job design

Yes, they did. However, I think there was an issue with the "Intelligent" part and some functionality that was documented was revoked or some such. Had an interesting conversation with someone over on ADN on that topic.

Oh, and thanks Kim for pointing out the fact that I still ask a question or two on occassion, while Ray and Ken have evolved beyond that need. [/kidding]

Kim, dont you think those questions were in bad taste?. I hate to see people utilizing this forum for "Interview" kind of questions rather than have a real time problem. We are lucky to have the best experts and guru's in this forum. Thanks for all the support. What does Ray, Ken, Chulett and other guru's have to say about this?

dsrules wrote:Kim, dont you think those questions were in bad taste?
...
What does Ray, Ken, Chulett and other guru's have to say about this?

I was letting my silence speak for itself, until you invoked my name. Yes these are interview questions. New user, first post, reads like a tech'ing cheat to me.

Okay, I was in a good mood. No more Mr. Nice Guy. From now on I will follow the unwritten rule. Sorry.

Even I have not been able to figure out the answer to Q13.

Maybe there's a case for providing partial answers, and alerting interviewers (via a premium service?) of some standard questions that can be asked, the standard partial answer that is published here, and some follow-up question to test the depth of the candidate's real knowledge.

On the topic of intelligent assitants, I've got one installed "New Data Migration Job", but I recall I once got a version 7.x installed which had intelligent assistants for slowly changing dimension jobs! I can't recall which client or machine I was on at the time time but I haven't been able to find it since. It created a sequence job an a bunch of server jobs for various types of SCD jobs.

Even I have not been able to figure out the answer to Q13.

Geez, Ray could not answer a question. I am .

Can YOU answer the OP's question 13?

(Arguably Q13 is the stunned silence from the interviewer resulting from the candidate's answers to the previous 12.)