Entering into Data warehousing

[know123]

Hi,

I have around 5 years of experience as a DBA and I'm now planning to move to OLAP/Data warehousing. I do NOT want to get into ETL side. I want to get into OLAP side. Would it be very difficult to enter this career? I'm now getting some training on OLAP related tools.

I would also like to know other people's opinion/advice on this

Thanks

 

[alpinestart]

let me know ow you expect to get into DW without ETL.

 

[johnherman]

OLAP (On Line Analytical Processing) tools are intended for end-users such as business analysts, managers, forecasters, etc., not for IS/IT folks. You could possibly add value by being able to build OLAP cubes from the database tables based on user specifications or your own knowledge of the business. This would involve learning dimensional modeling (star schemas) and you would need a fair to good knowledge of the business goals, strategy, KPI's, measures, dimensions, etc.

-------------------------
The trouble with doing something right the first time is that nobody appreciates how difficult it was - Steven Wright

 

[eyeswideclosed]

We don't use etl in our subsets because we haven't built our software on traditional relational databases, so we don't use etl to extract the data. The problem with etl is that every time you want to look at a particular set of data you have to rebuild the query.

 

[MDXer]

ETL does not mean you are only getting your data out of a relational database. in an ETL process the source of the extract can be any number of source data types flat files, databases, custom built datastores. if the data is comming from x into the DW then it is part of the ETL processes.

As for how to get into OLAP without doing ETL well that depends on the size of the company, and the ETL app just because someone is expert in MSAS doesn't mean they are expert in informatica. It is more the process you need to understand as a DW BI person. being a generalist in BI is a bit more difficult due to the broad nature of BI tools. Some tools are based around OLAP cubes such as Proclarity, AlphaBlox. Others support cubes and ROLAP reporting such as Cognos and Business Objects but are geared more towards reporting direct from the DW. WHile others don't use cubes at all like Microstrategy. If you want a Career in Business Intelligence you first need to find where your intrests and strengths are and go from there. A few classes won't do much good if you don't like what your doing.

In my opinion enjoying what you do is even more of a necessity in the BI word then other Dev enviroments given the complexity and size of projects it isn't something you can like today and not tomorrow and still hope to succeed.

Shoot Me! Shoot Me NOW!!!
- Daffy Duck

 

[know123]

Thanks for all your responses. I understand your point of view. At first I started to read about data warehousing just as a general reading but I started to develop more interest in this filed as I read more and more about it. Because of this only, I'm planning to get into Data wareousing field. There are again many areas in this field and it would be very difficult for me to enter into some of those without prior experience. That's the reason I mentioned about my backgroud and wanted to take advice from this forum members, as to which area/field in data warehousing would be relatively feasible to get into. I didn't want to get into ETL because my present job involves some ETL work eventhough we don't use any ETL tool. I'm not enjoying writing sql code and worrying about the nightly jobs that extract/load data. Let me now rephrase my question, If anyone wants to shift to data warehousing field (out of interest) what would be the more practical path to get an entry into any area of data warehousing. Then if that particular person has a strong database background then which area of data warehousing would be logically easy to get into? Once again thank you very much and I will look forward to your responses

 

[eyeswideclosed]

Don't think that "dw" is anything new either, people have been doing it for decades, it's just a new buzzword for something old, storing a massaged subset of data from the main repository into a secondary database, you can use "etl" to massage it if it is structured to a standard, or do it your self if it is not, which is what we do, you'll have to manually tweak the queries anyway more than likely, real life is messy.

 

[MDXer]

know123,

Unfortunately as someone entering the DW field your first stop will probably be the guy writing the SQL code to massage the data and load it into the proper tables. It all depends on where you are and the company enviroment.

I lead a BI team that includes report writers, cubes, dw pretty much the whole thing. Unless someone has at least my Skillset they will be the guy writing the code. Now I do run my team in such a fashion that the developers get the business reqs and the work with me on the architecture and the development plan and then they write the code. So pretty much my whole team has input into the design but their designs can be over ridden if the design is bad. and during hectic times it comes down to I say they do but those times are rare.

What area do you live in?

Shoot Me! Shoot Me NOW!!!
- Daffy Duck

 

[know123]

MDxer,

Thanks for taking time to answer my questions. I live in bay area. From your response, it looks like the only entry point to data warehousing for someone (with strong database background) new to this field, is ETL

 

[johnherman]

Well, there's the Data Modeling (Dimensional Modeling) aspect as well, but you will need a fair to good understanding of business terminology and process particular to your company.

-------------------------
The trouble with doing something right the first time is that nobody appreciates how difficult it was - Steven Wright