Hello!
I'm reading a Teo Lachev's book "Applied Microsoft Analysis Services 2005"
and one question appeared. Teo says, that
"Finally, note, that the dimension key doesn't have to form the bottom level
of a hererarchy" (p.85)
"Note that the leaf level of a multilevel hierarchy doesn't have to be
formed by the dimension key attribute..." (p.125)
But when I explore Microsoft Adwenture Works sample project, I see opposite
examples: look at Product Categories or Stock Level hierarchies of Product
dimension - there is "Product" dimension key at bottom level!
Can you explain to me, what approach is more correct and why? Maybe I
missed something, reading book on foreign language is too hard for me to
understand his ideas, or it is personal opinion of author that do not
corresponds to Microsoft's UDM ideology?

Thank you

I think Teo's point was that you do not need to include the key
attribute in all your hierarchies.

The key attribute defines the lowest level of granularity for the
dimension, but not every hierarchy in the dimension needs to go down to
the lowest level.

eg. In a date dimension, where the date is the key, you could just have
a year - month hierarchy.

--
Regards
Darren Gosbell [MCSD]
Blog: http://www.geekswithblogs.net/darrengosbell

In article <50631931-D510-4980-AA69-67E84D480656 (AT) microsoft (DOT) com>,
promethee (AT) community (DOT) nospam says...