Thanks for the quick response. I'll go and have a rethink of my tables. Cass xx
On Sat, 24 Jun 2006, Arek Kasprzyk wrote: > > On 21 Jun 2006, at 18:21, Caroline Johnston wrote: > > > Hi, > > > > I'm having problems getting placeholder attributes to work (see error > > message below). > > > > I've got a 'mirna' dataset which has a mature_name attribute from its > > mature dimension. > > > > I've also got seperate datasets for the predicted miRNA targets for > > each > > species. These are just single tables which have mature_name as an > > attribute. > > > > I've defined an exportable mature_name on mirna and importable > > mature_name > > on each of the targets datasets. This seems to be working: I can now > > select mirna as a second dataset in biomart_web. I can use the filters > > from the second dataset, so I can do something like 'give me all the > > target ensembl ids in this species which map to mature miRNAs derived > > from > > human miRNAs'. And I seem to be getting the right results. > > > > What I want to be able to do is something like: 'Give me the target > > ensembl ids and the associated miRNA database link in this species > > which > > map to mature miRNA derived from human miRNAs'. > > > > In other words, I want my result set to contain attributes from the > > target > > dataset and the mirna dataset. I tried to create placeholder > > attributes, > > like mirna.db_link in the target dataset, and they appear as selectable > > attributes on the output page, but if I try to select any of them I get > > the error below. Just selecting target attributes still works fine. > > > > Any idea what I've missed? Is there a way of viewing the query it's > > trying > > to run? > > > > Thanks, > > > > Cass. > > > > Hi Cass, > just to add that ... the easiest way to combine attributes from two > visible > datasets is simply to replicate data in dimension tables. It makes it > easier > for the user to interact with the data and the query performance is > better. > Ensembl mart uses this approach a lot for things like snps in gene > dataset > or homologies. We typically only use explicit joining of two datasets > when > one of them is remote, all other use cases are dealt with by introducing > extra dimension tables > > a. > > > > > > > > > > > <---------------- Error ---------------------------> > > > > SQL FAILED - LIKELY XML MISCONFIGURATION ERROR-------------------- > > EXCEPTION -------------------- DATE: 21-JUN-2006 MSG: returning undef > > ... > > missing filters from your importable? STACK > > BioMart::Configuration::FilterList::_toSQL > > /srv/biomart-web/biomart-plib/BioMart/Configuration/FilterList.pm:159 > > STACK BioMart::Configuration::BaseFilter::toSQL > > /srv/biomart-web/biomart-plib/BioMart/Configuration/BaseFilter.pm:290 > > STACK BioMart::Dataset::TableSet::_generateSQL > > /srv/biomart-web/biomart-plib/BioMart/Dataset/TableSet.pm:681 STACK > > BioMart::Dataset::TableSet::_fillAttributeTableWith > > /srv/biomart-web/biomart-plib/BioMart/Dataset/TableSet.pm:245 STACK > > BioMart::Dataset::TableSet::_getResultTable > > /srv/biomart-web/biomart-plib/BioMart/Dataset/TableSet.pm:152 STACK > > BioMart::DatasetI::getResultTable > > /srv/biomart-web/biomart-plib/BioMart/DatasetI.pm:573 STACK > > BioMart::QueryPlanner::_processPath > > /srv/biomart-web/biomart-plib/BioMart/QueryPlanner.pm:232 STACK > > BioMart::QueryPlanner::_processPath > > /srv/biomart-web/biomart-plib/BioMart/QueryPlanner.pm:346 STACK > > BioMart::QueryPlanner::_processPath > > /srv/biomart-web/biomart-plib/BioMart/QueryPlanner.pm:346 STACK > > BioMart::QueryPlanner::getResultTable > > /srv/biomart-web/biomart-plib/BioMart/QueryPlanner.pm:165 STACK (eval) > > /srv/biomart-web/perl/multi/martview:2993 STACK martview::print_results > > /srv/biomart-web/perl/multi/martview:2992 STACK > > Apache::ROOT::perl::biomart::martview::handler > > /srv/biomart-web/perl/multi/martview:227 STACK (eval) > > /usr/lib/perl5/site_perl/5.8.7/i586-linux-thread-multi/Apache/ > > Registry.pm:149 > > STACK Apache::Registry::handler > > /usr/lib/perl5/site_perl/5.8.7/i586-linux-thread-multi/Apache/ > > Registry.pm:149 > > STACK (eval) /dev/null:0 STACK toplevel /dev/null:0 > > ------------------------------------------- at > > /srv/biomart-web/biomart-plib/BioMart/Dataset/TableSet.pm line 681 > > > > > > > ------------------------------------------------------------------------ > ------- > 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 > ------------------------------------------------------------------------ > ------- > > > >
