From: Ugis S. <ug...@eb...> - 2002-01-18 13:20:44
|
Angel, > OK, all of these email conversations are running together for me so I want > to paraphrase some of what has happened thus far. > (Also, I think we are getting ahead of ourselves in talking about a data > warehouse, seeing as we do not have a MAGE schema on which to base it > on...) If we do a more or less straightforward mapping from MAGE-OM to database schema, then in order to run a simple query "give me all experiments that use arrays for ecoli" the number of tables that have to be joined (or traversed in some other way) is 8-10, depending on how superclass/subclass relations are mapped. That doesn't sound too good, does it? On the other hand, queries like that are needed here and now, so I don't think that discussions about data warehouses can be postponed for a year or so. If "data warehouse" is a too fancy term, we could use some other to describe means of making MAGE structures efficiently queryable. Cheers, Ugis |