Kimball Forum
Would you like to react to this message? Create an account in a few clicks or log in to continue.

Criteria for splitting fact tables

2 posters

Go down

Criteria for splitting fact tables Empty Criteria for splitting fact tables

Post  juanvg1972 Thu May 07, 2015 4:36 am


I have made a DWH matrix crossing measures with dimensions and now I have the doubt in some cases of creating a single table fact or varios fact tables.

I have 20 dims in my DW, also I have a group of similar measures (similar in business terms). I have 6 main measures in this group.
To that group of measures apply 10 dims of my DWH and all the measures of the group has 5 dims in commun. I can create a single fact table or two making a second division in my group of measures.

I would like to have some criteria in order to value that situation.
Here are my list of criteria to asses the situation, is ther something left? Can you give me more criteria?

- Physical space: with only one fact table, there will be a lot of nulls values.
- Grain of the tables
- Business meanning of the measures in order to group it.
- Loading process in the same table can cause bottleneck, difficult to parallel processing
- Queries from reporting tools to the same fact table, also it can cause bottleneck
- ...

Thanks in advance,


Posts : 25
Join date : 2015-05-05

Back to top Go down

Criteria for splitting fact tables Empty Re: Criteria for splitting fact tables

Post  ngalemmo Thu May 07, 2015 5:37 pm

The single criteria is the grain of the data. Never mix grains in fact tables. Any other criteria would have to do with dealing with different facts of the same grain.

Posts : 3000
Join date : 2009-05-15
Location : Los Angeles

Back to top Go down

Back to top

- Similar topics

Permissions in this forum:
You cannot reply to topics in this forum