In interest of having this information available to the community in case others also face such an issue, here is the explanation of this behavior and how to avoid it:

A scenario where a combination of events on the same chain (such as click and select, or populate and expand in this case) are one async and one full postback is not supported. This is done to avoid issues such cases can cause with event handlers. In reality there’s also rarely a reason to combine the two types of postbacks back to back.

For this specific scenario:

Perhaps part of the confusion comes from the multiple states of the AutoPostBackFlags – keep in mind the “On” value means full postback and there’s another enabled state that is “Async”. Combine that with the fact that Ajax is enabled by default for the Data tree (even without EnableAjaxViewState="true").

The events listing suggests the NodePopulate event should be used to handle when using IsEmptyParent, instead of NodeExpanded:

http://help.infragistics.com/Doc/ASPNET/2015.2?page=Infragistics4.Web.v15.2~Infragistics.Web.UI.NavigationControls.WebDataTree_members.html

So, as per the flags click and expand are full page reloads, but the populate is not described (therefore async)  and fires right before expand to load data and prevents the following event from triggering. Nodes with loaded children will naturally not require populate and their expand event will cause normal postback as expected. So to achieve loading in child data the simply avoid such event combinations and for the tree handle the populate first:

<ig:WebDataTree ID="WebDataTree1" runat="server" Height="600" Width="400px"

    CheckBoxMode="BiState" NodeIndent="18" EnableAjaxViewState="true"

    OnNodeClick="WebDataTree1_NodeClick" OnNodePopulate="WebDataTree1_NodePopulate">

        <AutoPostBackFlags NodeClick="On"/>

</ig:WebDataTree>

protected void WebDataTree1_NodePopulate(object sender, DataTreeNodeEventArgs e)

{

    DataTreeNode tn = new DataTreeNode();

    tn.Key = Convert.ToString(int.Parse(e.Node.Key) + 1);

    tn.Text = e.Node.Text + "_ChildNode";

    e.Node.Nodes.Add(tn); 

    e.Node.Expanded = true;

    e.Node.DataTree.DataBind();

}

I’m attaching the full sample and hope this is helpful.

Regards,

Damyan Petev

Associate Software Developer

Infragistics, Inc.

I believe I get your second case – you want to confirm the node expand action in a sense, right? Like a conditional manual load on demand. The population of node data is, however, strictly asynchronous. You could attempt to skip populating the node and force a postback after the population is "complete", but that would kind of go against the load on demand feature entirely.

The most straightforward solution I can think of is to cancel the client side NodePopulating event and save the node reference for later. Then go thought the user confirmation (show a dialog for example) and only after the user accepts, expand the node, which will go through the normal populating process as usual. This seems to make the most sense, so let me know if this is suitable for your case and I will try to provide more guidance on how to achieve that.

Regards,

Damyan Petev

Associate Software Developer

Infragistics, Inc.

I only assumed you can use something that can be displayed from a client API (like a Dialog Window) and frankly I would strongly suggest a similar approach. There’s rarely anything that can’t be converted from server rendering to client interaction anyway. This also assumes you can easily render if event should be canceled and verified from a general server condition. The case becomes a bit more complicated if you need server validation for each node separately (and even that can be done through simple endpoint request).

Cancelling the populating is easy enough:

<ClientEvents NodePopulating="WebDataTree1_NodePopulating"/>

<script type="text/javascript">

function WebDataTree1_NodePopulating(tree, e) {

e.set_cancel(true);

//handle validation, use e.getNode() to get a reference to the node

}

To force postback after that you can call in the standard __doPostback(). This article has a good explanation how to pass in arguments so your backend: http://www.codeproject.com/Articles/667531/doPostBack-function and you will need something similar to set the node information to validate and handle the server side, as there will be no event to do that for you.

Again I would only go through with the latter if the rest of the design is harder to change. I might be able to come up other solutions if you can provide a small sample with your general usage. Perhaps something other than manual load on demand can achieve the result you are looking for?

Lastly, if you think having an interruptible and fully synchronous Load on Demand for the Data Tree is important, perhaps you would consider submitting it on our Feature Suggestions site at http://ideas.infragistics.com/forums/192360-asp-net - that can be very helpful when product management decides what to add next to the product.

Regards,

Damyan Petev

Associate Software Developer

Infragistics, Inc.