On 12 Jan 2007, at 20:54, Marina Olhovsky wrote:
<snip>
However, MartBuilder does not place them in one dataset (the problem I
tried to solve from the beginning). It creates 3 distinct datasets:
- Experiments, which contains tables 'experiment' and 'isolate'
- Reagents, containing tables reagents, experiments, and a few
dimension tables related to reagents but unrelated to containers,
therefore not listed here
- Locations, containing tables wells, preps, containers and various
prep-related dimension tables
Should the tables listed be grouped into one dataset? Does it depend
on dimension tables not listed?
So many thanks for your help.
Hi Marina,
after a little investigation we have discovered that there was a
mistake in the way that transformation algorithm was implemented
ei, it was stopping one relation too short (link table) when creating a
dimension. This resulted in not sufficient number of tables
included in the transformation and this is likely to be related to the
problems you reported above. We have now corrected
that in the MBuilder for the new 0.6 release and although MBuilder
would still
attempt to create separate datasets in certain circumstances but the
datasets would be 'richer' allowing for connection between
them or even satisfying the queries within individual datasets.
However, the bad news is that this correction cannot be easily
implemented in 0.5 and will be left out from this release.
apologies,
a.
------------------------------------------------------------------------
-------
Arek Kasprzyk
EMBL-European Bioinformatics Institute.
Wellcome Trust Genome Campus, Hinxton,
Cambridge CB10 1SD, UK.
Tel: +44-(0)1223-494606
Fax: +44-(0)1223-494468
------------------------------------------------------------------------
-------
