Hot Swappable Dimension -> best practice for implementation of swapping logic?
2 posters
Page 1 of 1
Hot Swappable Dimension -> best practice for implementation of swapping logic?
Hi all,
could you help me with your experience regarding swapping the Hot Swappable Dimensions?
In Design Tip #16 is said:
"Hot swapping dimensions is straightforward in a standard relational database since the joins between tables can be specified at query time."
I would prefer to implement the swapping logic on database side, because my BI Tool struggles with specifying the join at query time.
Maybe you even have an idea for implementing it within Teradata?
Regards
Jochen
could you help me with your experience regarding swapping the Hot Swappable Dimensions?
In Design Tip #16 is said:
"Hot swapping dimensions is straightforward in a standard relational database since the joins between tables can be specified at query time."
I would prefer to implement the swapping logic on database side, because my BI Tool struggles with specifying the join at query time.
Maybe you even have an idea for implementing it within Teradata?
Regards
Jochen
Abild- Posts : 1
Join date : 2014-10-03
Re: Hot Swappable Dimension -> best practice for implementation of swapping logic?
I think the text in the design tip is a bit misleading as I think it only applies when querying your DB directly using SQL (using Oracle Sql Developer, MS SQL Studio, etc.) - whereas in most implementations a BI tool would be used to insulate the user from the underlying DB/SQL.
If you are using a BI tool then you would define the joins in your metadata layer when designing that layer - you wouldn't do it at query time as you would not expect your end-user to be aware of it happening, let alone how to force it to happen.
So looking at the 2nd example in the Design Tip (Retail Bank Account Types) you would create separate models for each account type in your metadata (with the appropriate joins defined) and then expose each model in its own presentation folder (or whatever the term is that your BI Tool uses)
If you are using a BI tool then you would define the joins in your metadata layer when designing that layer - you wouldn't do it at query time as you would not expect your end-user to be aware of it happening, let alone how to force it to happen.
So looking at the 2nd example in the Design Tip (Retail Bank Account Types) you would create separate models for each account type in your metadata (with the appropriate joins defined) and then expose each model in its own presentation folder (or whatever the term is that your BI Tool uses)
nick_white- Posts : 364
Join date : 2014-01-06
Location : London
Similar topics
» Fact to Dimension Join (Best Practice)
» Degenerate Dimension Implementation
» Junk Dimension Implementation
» dimension best practice!
» Best Practice: two Cube share one Dimension?
» Degenerate Dimension Implementation
» Junk Dimension Implementation
» dimension best practice!
» Best Practice: two Cube share one Dimension?
Page 1 of 1
Permissions in this forum:
You cannot reply to topics in this forum