Group & Bridge tables ... and a Surrogate Key on the Fact table?
2 posters
Page 1 of 1
Group & Bridge tables ... and a Surrogate Key on the Fact table?
I have a question about a design challenge that I’m facing regarding the use of a bridge table to handle a fact row associated with multiple rows of a dimension.
I have studied The Data Warehouse Toolkit with focus on the Multivalued Diagnosis Dimension section of Chapter 13 (p. 262-265).
I understand the concept and how to build the bridge table and the group dimension. My questions have to do with the following scenario:
The relationship between the group dimension and the fact table turns out to be 1:1
• Wouldn’t the group dim table be doing nothing more than providing a surrogate key for the fact table?
• And if that is so, would it be permissible to use a surrogate key on the fact table (eliminating the need for the group dim table) and allow the bridge table to serve as a standard cross-reference table?
• If that is ok, then what if the relationship (group to fact) is 98% 1:1 and only 2% 1:M? Would it be worth the additional query complexities of having the group dim table for the sake of 2%? (even the query on the load would be odious)
I have a second set of related questions…
• Can MS analytics cubes navigate the group/bridge correctly?
• How about with multiple group/bridge structures?
• How would it handle a surrogate key on the fact table and the XREF/Bridge table (no group dim)?
Thank you for your time.
David E. Martin
Tyson Foods, Inc.
IS Data Modeling
I have studied The Data Warehouse Toolkit with focus on the Multivalued Diagnosis Dimension section of Chapter 13 (p. 262-265).
I understand the concept and how to build the bridge table and the group dimension. My questions have to do with the following scenario:
The relationship between the group dimension and the fact table turns out to be 1:1
• Wouldn’t the group dim table be doing nothing more than providing a surrogate key for the fact table?
• And if that is so, would it be permissible to use a surrogate key on the fact table (eliminating the need for the group dim table) and allow the bridge table to serve as a standard cross-reference table?
• If that is ok, then what if the relationship (group to fact) is 98% 1:1 and only 2% 1:M? Would it be worth the additional query complexities of having the group dim table for the sake of 2%? (even the query on the load would be odious)
I have a second set of related questions…
• Can MS analytics cubes navigate the group/bridge correctly?
• How about with multiple group/bridge structures?
• How would it handle a surrogate key on the fact table and the XREF/Bridge table (no group dim)?
Thank you for your time.
David E. Martin
Tyson Foods, Inc.
IS Data Modeling
David Martin- Posts : 1
Join date : 2009-03-25
Location : NW Arkansas
Re: Group & Bridge tables ... and a Surrogate Key on the Fact table?
It would be helpful if you were more specific as to what the group is supposed to be.
If, as you say, the group has a 1:1 relationship with the fact table, then it doesn't make a lot of sense to have a group. But the only thing I can think of that would fit a situation like that would be a MySpace friends list...
If, as you say, the group has a 1:1 relationship with the fact table, then it doesn't make a lot of sense to have a group. But the only thing I can think of that would fit a situation like that would be a MySpace friends list...
Similar topics
» Bridge tables versus massive junk dimensions
» Bridge Table - To simply group a dimension in more than one way?
» Type 2 Fact table - Durable surrogate key?
» separate fact table/different grain - do I need a bridge table
» Fact surrogate key as foreign key in another fact table
» Bridge Table - To simply group a dimension in more than one way?
» Type 2 Fact table - Durable surrogate key?
» separate fact table/different grain - do I need a bridge table
» Fact surrogate key as foreign key in another fact table
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum