The Vertica Database Forums

Hi,
I am new to Data modeling side and i am trained on modeling techniques.

Are these techniques are specific to a traditional RDBMS?

I have some knowledge on Vertica on how the data is stored... the projections, compressing and encoding techniques.

is there anything that i may have to consider for Data Modeling on a Vertica DB? since its a columnar!
is there any specific documents i need to go through for that?


Thanks,
Jagadeesh.

Hi Jagadeesh,

There are at least two kinds of data model techniques, one for transactional databases and the second technique is for Business Intelligence. In the transactional data model we work with normalization techniques, and with the BI models is usual to work with "Cubes" and the database is not exactly normalized. In BI data models the "Fact and Dimension" concepts are usual.

And you need to use the reporting requirements for your data model and keep focused the data model in the analysis and reporting requirements.

In the transactional databases we need a few resources for many users (sometimes millions) and in the BI or Big data Databases as Vertica, we will require many resources for a few users.

You need to read about "Cubes data modeling" or "Analytic Databases and Dimensional Design", you can read the book: Star Schema: The Complete Reference, this is just a sample.

Thanks,

Francisco