Archives of the TeradataForum

Message Posted: Wed, 23 Feb 2011 @ 13:15:10 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

I recently did rather similar, that being to Create the Dimensions from the Fact Table although, if I understand you correctly, my implementation was slightly different to yours ....

my implementation was to create the Dimensions straight from the heavily denormalized Columns on the FACT ROWS, such as Country Name.

I created Aggregate Sparse Join Indexes on these Columns and then presented a Dimension View over the top of the Join Indexes.

That way, we didn't even need to separately create Dimensions - they were there in the data (albeit simple Code and Description ones) !

Clearly, if you've got a very rich set of Dimension Attributes then our approach would be inappropriate.

Anyway, in your case, however, I suspect you've got Dimensions as real and SEPARATE entries on the SAME Table as the Fact.

That being the case (and I won't say "I wouldn't do that"), as long as you can get to those Dimension entries quickly (and I'm alluding to Aggregate Join Indexes, 'picking' them out), then you should see what I saw, which is Teradata Joining the Join Index to the Base Table Fact !

All very clever really.

Regards

David Clough
Database Developer
Database Design Group

	<Prev	Next>		<<First	<Prev	Next>	Last>>