It seems that an error is raised when a deleted node is flushed when
post_update flag is set (See attached test case). Working with rev 1855.
After the second flush, the expected sql output should be:
[engine]: UPDATE node SET prev_sibling_id=? WHERE node.id = ?
[engine]: [2, 4]
[engine]: UPDATE
I should note that the relationship you are actually trying to set up
includes a circular relationship between two individual rows. by
default this is impossible to accomplish with only INSERT statements
and not violate foreign key constraints. to break this circular
relation you have to
you need it on 'parent' also
On Sep 5, 2006, at 12:55 PM, Nick Joyce wrote:
> I had specified the foriegnkeys for previous/next_sibling to
> node_table.c.id, but it appeared to make no difference, so I removed
> them in case is was casing a problem.
>
> Setting these relations to your suggestion
I had specified the foriegnkeys for previous/next_sibling to
node_table.c.id, but it appeared to make no difference, so I removed
them in case is was casing a problem.
Setting these relations to your suggestion still fires the same error.
Michael Bayer wrote:
> your relationships are not being se
your relationships are not being set up correctly, since you haven't
specified a "foreignkey" parameter on any of them. note that the
"root" relationship in the documentation example has a "foreignkey"
set on it.
when you say:
> node_table.c.parent_id==node_table.c.id
the "foreignkey" is
I am attempting to extend the self referential example in the docs
http://www.sqlalchemy.org/docs/index.myt?paged=no#advdatamapping_circular
by adding properties 'prev_sibling' and 'next_sibling'. (See attached
script)
When run, the following exception is raised:
Traceback (most recent call last)
6 matches
Mail list logo
