Well, not 'change' as in that I need to check for changes every once in a while, but more like if I have a saved state from an earlier usage of the treetable (i.e. I selected a node). If I then 'editi' this selection I was thinking that

<q:singleNodeSelection nodeData="#{activityAgentTree.activityNode}" /> would at least expand the treetable up unto the point of selection and selecting that node. It could also be that my getActivityNode() isn't properly implemented. I have a generic Tree implementation that wraps the data of the node. Upon selecting a node I only save this data of course, so when I want to edit this data I have to do something along the lines of:

Activity activity = someManager.getActivity();

return new JobActivityNode<Activity>(activity);

I was assuming that because hashCode() and equals() are correctly implemented the tree would now which node to select + expand. But maybe this is not the case.

PS: is is just me or do I miss a way to 'code-format' my code? Not to mention the rest of the messy formatting.

Thank for sticking with me

Indeed the saving works like a charm, however I was under the impression that binding this to a property would also automagically select the right node again if the binding property had a non-null value. I'm not terribly fluent with JSF yet, but if I use a value along the lines of

<h:inputText value="#{manager.property}" ></h:inputText>

for my inputText and this property gives me a non-null value (i.e. I edit the form) it will present me with this value in the form. Therefore I thought that populating the nodeData with a saved value:

public TreeNode getActivityNode() {
  return activityNode;
}

would present me with a tree that had the correct selection, and the correct expanded state. However this does not work for me, but maybe this is by design.

If the above doesn't work automatically, it boils down to this: I will have to retrieve the information from the database, look it up in my tree-structure (which only exists on the 'UI' side), and then call setNodeExpanded( ... ) on every parent of that node?

The value property of the datatable? Am I really missing something or do you mean that it should be bound to the same type as an element in the list that is given to the nodeChildren attribute of the dynamicTreeStructure element? In that case TreeNode is correct, because those are actually the elements in that list.

I must go on with some other stuff for now, so I'll try to figure some things out before my evaluation ends, but I'll be back in a few days to this thread I guess.