I had thought about that, but it would require re-writing all my other code that depends on having all the nodes populated. It's kind of like putting half of a table in a grid and then trying to do a seek on that half of a table to find something that may be in the other half. I really need to have all the data available at the same time. The user probably won't look at it all at once, but it needs to be there for behind the scenes stuff.

Thanks,

Michelle


>Hi Michelle,
>
>Is it really necessary to have all the nodes populated?
>
>It's pretty unlikely that ny one user would want to see >, say, 100 in any one viewing. Even if you increase that by a factor of 100, you're still well under the limit.
>
>You could load the top level nodes and put a place-holder node under each of those top level nodes that has children. When a node is expanded, check to see whether it has the placeholder node and if so, remove the placeholder node and load the child nodes, along with a placholder node under each of the child nodes that has children. Repeat ad-infinitum.
>
>If you need to, keep a counter of the number of nodes currently popluated and take some action if it gets towards the 64K limit.
>
>This technique could have the added effect of making your application appear much faster as only the data that are required are loaded into the tree.
>
>Cheers,
>
>Andrew