django-mptt get_descendants for a list of nodes

Later versions of mptt already have this function built into the object manager. So the solution to this is just as follows:

Node.objects.get_queryset_descendants(my_queryset, include_self=False)

Great thanks to Craig de Stigter answered my question on django-mptt-dev group, in case anybody need it I am kindly reposting his solution from http://groups.google.com/group/django-mptt-dev/browse_thread/thread/637c8b2fe816304d

   from django.db.models import Q    import operator    def get_queryset_descendants(nodes, include_self=False):        if not nodes:            return Node.tree.none()        filters = []        for n in nodes:            lft, rght = n.lft, n.rght            if include_self:                lft -=1                rght += 1            filters.append(Q(tree_id=n.tree_id, lft__gt=lft, rght__lt=rght))        q = reduce(operator.or_, filters)        return Node.tree.filter(q) 

Example Node tree:

T1 ---T1.1 ---T1.2 T2 T3 ---T3.3 ------T3.3.3 

Example usage:

   >> some_nodes = [<Node: T1>, <Node: T2>, <Node: T3>]  # QureySet   >> print get_queryset_descendants(some_nodes)   [<Node: T1.1>, <Node: T1.2>, <Node: T3.3>, <Node: T3.3.3>]    >> print get_queryset_descendants(some_nodes, include_self=True)   [<Node: T1>, <Node: T1.1>, <Node: T1.2>, <Node: T2>, <Node: T3>, <Node: T3.3>, <Node: T3.3.3>] 

Django mptt uses the modified pre-order tree traversal method as described in the MySQL Managing Hierarchical Data document.

It has the following query for returning all nodes in a tree below a certain node:

SELECT node.nameFROM nested_category AS node, nested_category AS parentWHERE node.lft BETWEEN parent.lft AND parent.rgt    AND parent.name = 'ELECTRONICS'ORDER BY node.lft;

The secret is the parent.lft and parent.rgt numbers, all children will have a node.lft value between the two.

Obviously that example presupposes only one parent and that you need to use the parent name to find the parent. As you have the parent node data already you can do something like the following:

SELECT node.idFROM node_tableWHERE node.lft BETWEEN parent[0].lft AND parent[0].rgt    OR node.lft BETWEEN parent[1].lft AND parent[1].rgt

I'll leave it as an exercise for you as to how to generate a seperate BETWEEN clause for each parent node (hint, " AND ".join)

Alternatively you can use a the range generator on each parent to obtain all the values between each parent's lft and rgt values inclusive. That then lets you use a giant IN statement rather than lots of BETWEEN clauses.

Combine either of the above with a RawQueryset and you will get the models back.