Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
Vijaykumar_Patil
Posting Yak Master
121 Posts |
 Posted - 2008-06-05 : 08:01:37
|
| Hi All,I am trying to understand, when would I do a vertical partition in a Dimensional Data Warehouse ? What are the things I need to consider, before I take the decision?Necessity is the mother of all inventions! |
|
|
Vijaykumar_Patil
Posting Yak Master
121 Posts |
Posted - 2008-06-05 : 09:15:59
|
| I think .. I am asking in a wrong way.. I think I am looking for snowflake schema. As to when to do it and what are the things I need to consider to make decision to use snowflake schema.Necessity is the mother of all inventions! |
 |
|
|
rmiao
Master Smack Fu Yak Hacker
7266 Posts |
Posted - 2008-06-05 : 22:54:19
|
| Don't do it because of performance. |
|
|
|
Lumbago
Norsk Yak Master
3271 Posts |
Posted - 2008-06-06 : 02:42:20
|
| If you don't have large amounts of data a snowflake design could be defended and work somewhat well for you, but once the amount of data increases you will face performance problems just like reporting on a traditional relational database model. Try creating star schemas instead and remember that there is always the possibility of having several different fact tables using the same dimensions.--Lumbago |
|
|
|
Vijaykumar_Patil
Posting Yak Master
121 Posts |
Posted - 2008-06-06 : 07:21:28
|
| thank you guys I understand that part very well. I am trying make a check list before I split a dimension. If at all I need to split it. So link or a few points would be great.Necessity is the mother of all inventions! |
|
|
|
blindman
Master Smack Fu Yak Hacker
2365 Posts |
Posted - 2008-06-06 : 10:45:05
|
| Star schemas are for datamarts, not data warehouses. Keep your warehouse in snowflake, or better yet, relational form. Then add preaggregated tables or spin off subject-area datamarts for performance as the need arises.e4 d5 xd5 Nf6 |
|
|
|
manstein
Starting Member
1 Post |
Posted - 2008-08-20 : 10:57:09
|
Its a game  Bb5quote:
Originally posted by blindman
Star schemas are for datamarts, not data warehouses. Keep your warehouse in snowflake, or better yet, relational form. Then add preaggregated tables or spin off subject-area datamarts for performance as the need arises.e4 d5 xd5 Nf6
Its gonna be a long, long way to happy |
|
|
|
blindman
Master Smack Fu Yak Hacker
2365 Posts |
Posted - 2008-08-20 : 13:20:55
|
| Took me a moment to figure out what you were talking about. Ha!Boycott Beijing Olympics 2008 |
|
|
|
|
|
|
When to make a Vertical partition in Data Warehous