Dimentional Modelling Insight
2 posters
Page 1 of 1
Dimentional Modelling Insight
Hello Guys, I work in a claim management company and we are implementing our first DW, I am trying to de-normalized various tables to make a single dimension. My questions
1. I have 14 dimensions already and still need more dimensions, how many dimension can a data warehouse have?
2. I have 12 tables that is associated with a claim table all using the business key to link each other to some extent while few using primary keys, is it possible to join all 12 tables into one large table, in this case, can I have a snowflake dimension or join all..if the later is the case, wont it be a problem to have more than 25 columns in a single dimension?
3. T-SQL vs SSIS slowly changing dimension: which is best to use as I am not good in query writing but the test I did on ssis slowly changing dimension wizard is too slow
I will be glad if someone can lecture me and sorry all for the pains in reading
1. I have 14 dimensions already and still need more dimensions, how many dimension can a data warehouse have?
2. I have 12 tables that is associated with a claim table all using the business key to link each other to some extent while few using primary keys, is it possible to join all 12 tables into one large table, in this case, can I have a snowflake dimension or join all..if the later is the case, wont it be a problem to have more than 25 columns in a single dimension?
3. T-SQL vs SSIS slowly changing dimension: which is best to use as I am not good in query writing but the test I did on ssis slowly changing dimension wizard is too slow
I will be glad if someone can lecture me and sorry all for the pains in reading
Prince101- Posts : 2
Join date : 2013-02-05
Re: Dimentional Modelling Insight
I work in the insurance field. Claim fact tables can be quite leggy - customer info, client info, service provider info, product, dates, adjudication info, etc. You figure at least 4 date fields on the claim - incident date (or service date), date the claim was received, adjudication date, paid date. Potentially 4 Address fields, 4 zip code related fields. That's 18 dimensions and you haven't adjudicated the claim.
You kind of have to play around with it. Try IDying smaller dimensions that could be combined into a junk dimension - but sometimes combining 2 smaller dimensions can explcause the number of rows to explode.
You kind of have to play around with it. Try IDying smaller dimensions that could be combined into a junk dimension - but sometimes combining 2 smaller dimensions can explcause the number of rows to explode.
Jeff Smith- Posts : 471
Join date : 2009-02-03
Similar topics
» What's the best dimentional design for this situation?
» Dimensional Modelling compared to Relational (3NF +) Modelling for DSS
» Different time zones
» Handling different Time Zones
» Modelling KPI values
» Dimensional Modelling compared to Relational (3NF +) Modelling for DSS
» Different time zones
» Handling different Time Zones
» Modelling KPI values
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum