On 5 Jan 2007, at 15:33, David Croft wrote:
Hi Arek,
he difference between two tables seem to be 5 attributes from the
Reaction table:
| inferredProt | text
| YES | | NULL | |
| maxHomologues | text
| YES | | NULL |
| reverseReaction | int(10)
unsigned | YES | | NULL | |
| reaction_stableIdentifier_key | int(10)
| YES | MUL | NULL | |
| totalProt
the reaction_stableidentifier_key will be ignored because it ends
with '_key'
are you saying that MEditor is ignoring the other 4 and you don't get
any attributes
or fitlers for them?
Yes. What's more, although the tableConstraint for the attributes is
'main', the
key is 'pathway_stableidentifier_key' for *all* attributes.
Would it help if I removed the attributes that differ between the two
main
tables?
nope, a subclass table is allowed to have as many extra attributes as
it wants
as long as it replicates everything from the superclass one. We need a
test
MEditor run with your tables to see what's happening
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
------------------------------------------------------------------------
-------
