topic Re: SAS Servers :: Data Warehousing in SAS Data Management

SAS Servers :: Data Warehousing

HigherPurp0se — Thu, 21 Jun 2018 03:32:21 GMT

Hello, all.

I am a new student learning data science, and have just been introduced to SAS SUE.

In my next course I will be learning about data warehouses.

Does SAS Institute have any data warehouses in a test environment available to us to experiment with and learn on?

What other SAS software toolset must I learn in order to build a data warehouse on a cloud resource such as Azure, AWS or Hadoop?

Thank you for your consideration.

Re: SAS Servers :: Data Warehousing

Patrick — Thu, 21 Jun 2018 06:09:15 GMT

Data Warehouses are mainly about how to organize your data so they are easy to maintain and access(query). On this level Data Warehouses are agnostic to technology.

Data Warehouses come in many different shapes and sizes so it will depend on what your course will focus on.

But if you're after some SAS file based sample data then Cynthia's answer here might guide you.

https://communities.sas.com/t5/General-SAS-Programming/How-to-access-sample-orion-star-data-files-used-in-the-course/td-p/211165

Re: SAS Servers :: Data Warehousing

LinusH — Thu, 21 Jun 2018 06:13:24 GMT

I'm not aware of any SAS sample environments that should reflect a data warehouse.
The main SW in SAS arsenal specific for DW IMO is Data Integration Studio. And that is unfortunately not available in any of the learning editions of SAS. That said, DW is a concept, not SW. Anyone with a data store and scripting/programming environment could build a data warehouse.
Not sure if there is any cloud specific SAS trainings, but knowledge about SAS/ACCESS should come in handy.

Re: SAS Servers :: Data Warehousing

HigherPurp0se — Thu, 21 Jun 2018 12:53:02 GMT

Thank you, Patrick.

Re: SAS Servers :: Data Warehousing

HigherPurp0se — Thu, 21 Jun 2018 12:53:30 GMT

Thank you, LinusH.

Re: SAS Servers :: Data Warehousing

AngusLooney — Thu, 21 Jun 2018 23:07:52 GMT

DI Studio is a tool for building all the processes for ingesting data into (and out of) a DW, and there's nothing inherent in it that takes any particular form or design of DW into account.

I think it's best to think of DIS as an ETL tool - "Extract Transform Load" (as well as some of the conceptual variants of paradigm like ELT.)

I would separate "Data Warehouse design" from "ETL design".

There are definitely a number of "patterns" that would come to mind, in thinking of ETL design - but that's a subject for a longer post.

What I would say is that the primary considerations are about long term cost of ownership and operation of the ETL processes, rather than the minutiae of "DW design".