Re: Tree structure table normalization problem (do I need atrigger?)

On Tue, 19 Dec 2000, Frank Joerdens wrote:

> Tulassay Zsolt wrote:
> [ . . . ]
> > The SQL stuff of that nested set structure is fairly easy, I wrote some
> > quick'n'dirty plpgsql functions that will do inserts, updates, deletes
> > from the tree, display level number etc.
>
> What scared me about it in particular was one scenario where you try to delete a subtree.
> This would normally leave a gap, since the structure is based on the worm's ability to get
> from one node to the next with an increment of just 1. Once you had a subtree deleted,
> you'd either would have to have the worm leap-frog (a leaping frog-worm?!! :)) the gap or
> update an entire half of the tree to close it . . . then my brain started to hurt and I
> gave up.
>
that's exactly why i wrote a deletion function which updates the nodes'
"left" and "right" values to close the gap.
sure it causes a bit of overhead to update many nodes, but if you
don't delete subtrees all the time, you can live with this.

Zsolt Tulassay