>thnx all for your help, i have made the node with the help of cetins example. Starting with a dbf of 1600000 records in combination of other files i managed to make one cursor that contains all info i need to build my treeview. my final cursor contains 66130 records. Building the cursor takes the pc max 10 seconds. putting the info 66130 records takes me 50 seconds.
>
>this is my code :
>
>SCAN
> IF empty(mercedes_treeview.parentID)
> .Nodes.add(,tvwFirst,alltrim(mercedes_treeview.nodeID),ALLTRIM(mercedes_treeview.nodeText))
> ELSE
> .Nodes.add(alltrim(mercedes_treeview.parentID),tvwChild,alltrim(mercedes_treeview.nodeID),ALLTRIM(mercedes_treeview.nodeText))
> ENDIF
>ENDSCAN
>
>Does somebody knows how i can speed up the building of the treeview ?
>
>thnx

Kurt,
As I said in my original post, read June 2001 issue of UT magazine and download code from there. In that I showed a near 'ligthning fast' loading with an 'on demand' logic. Sample populates the nodes from folders as windows ecplorer does but easy to implement with a different node provider (and actually later I created a nodeProviderBase class which separates the node provider from UI - but finding it on my disk now is not easy for me).
Cetin