It's missing something, presumably you want to return the traversed FKs (or
their .q expression) in addition to the classes found in the search (building
the expressions as X.id == Y.id will not get the results intended, you want
X.fkToYID == Y.id). I'd probably let .q handle more of the work in
BuildJoinExpression too, so for instance:
(in BFS:)
pfkeys = fkeys + [(cls, col)]
(in BJE:)
exprs = []
prevTable = baseTable
for (table, col) in tables:
# col in prevTable references table
left = getattr(prevTable.q, col.name)
rght = table.q.id
exprs.append(left == rght)
prevTable = table
return AND(*exprs)
---
Using .q notation for traversing Joins and FKs is in trunk, so the non-automated
way to address your problem is much simpler now:
Foo.select(AND(Foo.q.bar, Bar.q.blah), orderBy=Blah.q.sortme)
or equivalently if you define the matching SQL*Joins:
Foo.select(AND(Bar.q.foos, Blah.q.bars), orderBy=Blah.q.sortme)
Oleg poked me again about documentation, so I'm getting on it.
Can't come up with a straightforward cyclical example that breaks BFS using only
FKs, if this were extended to other joins you'd also presumably be more pressed
to deal with hinting for disambiguation of multiple paths.
Nice work,
- Luke
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
sqlobject-discuss mailing list
sqlobject-discuss@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sqlobject-discuss
