On 6 Feb 2007, at 14:00, David Withers wrote:
How can I determine the ordering of query results when using
martservice? I thought that the order was determianed by the order of
the attributes in the query. However, in the following query:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE Query>
<Query virtualSchemaName="default" count="0" softwareVersion="0.5"
requestId="taverna">
<Dataset name="uniprot">
<Attribute name="sptr_ac" />
<Attribute name="uniprot_gene_sequence" />
<Filter name="superregnum_name" value="archaea" />
<Filter name="proteome_name" value="Aeropyrum pernix" />
</Dataset>
</Query>
the results are returned in the order 'uniprot_gene_sequence, sptr_ac'.
Hi David,
We have now done a little investigation here and unfortunately this is
going to remain an apparent
exception to the rule for a little while longer. The ordering of the
result columns
as you correctly point out is determined by the dataset order first (if
there is more than one of course)
and then by ordering attributes within individual datasets. Since in
most cases you have control
over both(the datasets need to be visible and that's what you deal with
most of the time)
you can always determine the order of the results.
However, we do have a concept of an invisible dataset in mart. This
dataset is added implicitly
during query execution on the server side and although it does actually
follow the same rules as
ordering of visible datasets ei dataset order first, attribute order
second, the user has no control over
explicit ordering of datasets in this case and this results in the
inconsistency as reported in your example.
Fortunately all other invisible datasets either have no attributes or
there are other alternative arrangements
in place so they can be used 'safely'. The only exception is
GenomicSequence so each time you
request an attribute from genomic sequence, the order will be rigidly
defined. I am afraid for now
you will have to live with that and simply treat a few of these
attributes differently from the others. We do
accept that this inconsistency is annoying and In the long run we'll
try to come up with a more user friendly arrangement
BTW, if you need this for your FASTA formatting purposes, you can
always request FASTA format
from martservice and this will always come formatted correctly
a.
David.
--
David Withers
School of Computer Science, University of Manchester,
Oxford Road, Manchester, M13 9PL, UK.
Tel: +44(0)161 275 0145
------------------------------------------------------------------------
-------
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
------------------------------------------------------------------------
-------
