Talk:Star schema

For the sake of those who aren't theoretical database experts, it'd be nice to have a counterexample of how you'd represent the same data in a traditional relational schema and how queries against the two schemas would differ. As someone who's learned SQL the hard way and hasn't had any formal database training, it's difficult to see what the big deal is. (Not saying there isn't one, you understand. Just that it would be nice if the article illustrated what it is.) --Chronodm 16:33, 31 October 2005 (UTC)

What the big deal is... good question. Ralph Kimball seems to be one of the key players in convincing many people that its a big deal. http://www.noumenal.com/marc/schema.html http://www.oreview.com/9602burl.htm Some day, I hope enough people realize it's not the panacea they were led to believe it was. Instead, think about the complex cause and effect process you are trying to model. In most cases, the truly useful big questions aren't answerable just with the corporate data being compiled to track sales and costs (regardless whether it's in 5th Normal Form or a Star Schema). Then go find a way to approximately get that data, and leave the banal stuff (including OLAP and Star Schemas) for the soon to be outsourced projects.

sql examples
Please, someone give the sql examples an introduction so their relevance is obvious to people, otherwise they are worthless and should be (shall be?) deleted. 75.73.52.72 03:29, 29 December 2006 (UTC)dave


 * They aren't worthless without a description and should not be removed. I do agree that they need such a description or that we need some better examples, but I disagree strongly with the assertion that they should be removed without one. --Nachtrabe 16:49, 6 February 2007 (UTC)

There's no sql example in this page. Try to load new one.


 * Is there any reason the current example is written without the use of inner joins syntax? Or are we just going old school? Root4(one) 20:47, 29 October 2007 (U

Could anyone draw a simple diagram illustrating this? My understanding is just that this is one main table with a bunch of foreign keys but I'm not familiar enough with this to be sure. 216.165.132.250 20:04, 25 September 2007 (UTC)


 * Surely we can put together a simple star schema database diagram? If not a physical diagram, at least a simple logical one?  If no one with actual graphics talent takes me up on this in the next week or so I'll make an attempt at it.  I have to warn you though, graphics are not my forte.  Thanks SqlPac (talk) 05:15, 24 June 2008 (UTC)


 * Added a simple image SqlPac (talk) 17:49, 27 June 2008 (UTC)

Now there is a difference between the schema names in the description and the code and the ones used in the diagram. Could anyone please change that? Are periods allowed in table names without quotation of the name? If not the names should probably be changed to use underscores instead of periods. --95.114.79.51 (talk) 11:02, 14 August 2010 (UTC)

Just a question ...
for the image on this page to be compatible with the text, should not all the table names be consistent? For example, the image shows "Fact_Sales" while the text, and the sql example, both use "Fact.Sales" ? KingGeezer (talk) 14:25, 11 March 2011 (UTC)

Fixed. &quot;Six by nine. Forty two.&quot; (talk) 07:44, 19 June 2011 (UTC)

More examples
We should flesh this out with a few tables as a supporting example. Maybe they could be the same tables used in the SQL query example? SqlPac (talk) 20:31, 27 June 2008 (UTC)
 * In fact, it would be a good idea to coordinate the example with the snowflake schema article so that we could give two consistent but contrasting views of the same schema. We could probably carry this concept over to the appropriate article on normalized relational models as well. SqlPac (talk) 20:36, 27 June 2008 (UTC)
 * Added a diagram of an actual star schema that reflects the given code sample. Changed table names to make them more descriptive, and modified code slightly.  I think we should port a similar example over to the snowflake schema article. SqlPac (talk) 21:10, 27 June 2008 (UTC)

Minor edit
Removed the following: "On an Entity-Relationship (ER) diagram, fact tables have few distinct columns, while dimension tables have a large number of columns. However, most of the storage is used by the fact table." This sentence is false, so I have removed it. SqlPac (talk) 01:14, 28 February 2013 (UTC)

Rewrite
Just rewrote a large chunk of this one to try to make it flow a little better and provide some structure to it. SqlPac (talk) 02:27, 28 February 2013 (UTC)

Remove the lines in the Disadvantages Section without any citations
This needs to be done because Star schemas that have any many to many relationships can be solved via a factless fact table, which can establish the connections between these many to many relationships quickly, which is how it is done in OLTP systems. — Preceding unsigned comment added by 54.164.135.164 (talk) 00:24, 26 June 2017 (UTC)

Suggestion for adding example in Life Sciences
I propose adding a link to an example of a well-known open-source star schema data warehouse in the Life Sciences, that I've contributed to and is currently an orphan page: https://en.wikipedia.org/wiki/TranSMART --Wardweistra (talk) 10:55, 15 January 2019 (UTC)