Comments on: Database vs. Data Warehouse

By: Alex

Alex — Sun, 16 Oct 2011 01:18:24 +0000

Hi Nishith & Patrick (comment #63),

Wonderful wealth of knowledge you have given to me! Much appreciated and bookmarked.

Cheers
Alex

By: niks

niks — Sun, 09 Oct 2011 16:26:13 +0000

Hey, i have 1 question.

Can a company have more than 1 datawarehouse ?

By: musumba banda

musumba banda — Tue, 04 Oct 2011 20:54:02 +0000

Interesting article, it makes understanding data base and data ware house very easy.

By: Patrick

Patrick — Tue, 20 Sep 2011 22:57:22 +0000

Data warehouses and databases do only have a small semantic difference when using relational products such as MySQL and Oracle. There truly are no differences in technology when implementing a DW or database. You optimize for reads typically using a star or snowflake schema in a DW, and a database uses an OLTP structure using 3NF schema which is optimized for writes and minimally optimized for reads. This is true even applying to column store databases such as Infobright, Vertica, Vectorwise and LucidDB which offer significant read performances on star/snowflake schemas.

However DWs take on a whole new meaning when you consider proprietary technologies like Netezza, Aster Data which are a combination of hardware and software. These are not databases. These are appliances that don’t even operate on a relational level.

Taking it further with open source, you have very scalable technologies like Hadoop with Hive and HBase which provide interfaces allowing for relational-like queries, but operate on the principles of MapReduce which are massively scalable on commodity hardware.

Data warehouses do in fact have significant and measurable differences when using the appropriate technologies. MIS/IT departments building DWs on MySQL or Oracle quickly find themselves in search of alternative technologies simply because while they can serve the purpose of a DW, they simply weren’t designed for it. I’d highly recommend doing some reading on MapReduce and Hadoop and you can see some very real-world examples where terabyte sized DWs are not only functional, but solving very real problems of Big Data, performance and reliability.

By: Daryl

Daryl — Mon, 19 Sep 2011 16:11:41 +0000

Interesting article, but it left out one other compelling reason for a separate data warehouse repository for analytics: if you have a need to report based on aggregated information housed in disparate databases/applications.

By: kim

kim — Thu, 28 Jul 2011 14:51:04 +0000

thanks! appreciate the great summary, esp for newbies.

By: Gunjan

Gunjan — Fri, 22 Jul 2011 10:42:24 +0000

Thanks a lot Nishith.. Its really helpful

By: Govind Sidhu

Govind Sidhu — Fri, 24 Jun 2011 07:30:30 +0000

nice article

By: Ram

Ram — Wed, 08 Jun 2011 06:18:39 +0000

thanks nitesh. it’s very good

By: David

David — Mon, 30 Aug 2010 08:34:35 +0000

It’s good and know more about data base and data ware house and thank’s for my presenttation