Hi Leandro Based on your initial input MBuilder will choose certain tables to include in the dataset according to the rules used in the transformation algorithm. This is to make sure that the result of the tranformation will comply with the reversed star schema spec. In short the rules are as follows:
1. Start with the source table(s) provided by the user and merge all the tables along the longest path satisfying 1:1 and n:1 relationships to create main table(s) 2. Create dimensions following the rules as above starting at the point that where it stopped merges for the main tables. No other tables will be included in this dataset. Hope this was helpful a On 2010-04-22, at 7:19, "Leandro Hermida" <[email protected]> wrote: > Hi everyone, > > I have a most likely novice question about MartBuilder, I have a > small source normalized schema describing protein models and > relationships to other entities. I want MartBuilder to build just > one dataset around the protein models with all the other entity > relationships included in as many dimensions as it needs. > MartBuilder just ignores certain tables in the source schema and > doesn't seem to make one dataset covering everything. > > Are there some rules in BioMart that prevent you from taking a > certain entity in a schema and designing one dataset with all the > related data? > > thanks for the help, > Leandro >
