Loadondemand property | Infragistics Forums

660

Loadondemand property

chris posted over 15 years ago

Also, I'm wondering about the loadondemand property. If it is automatic, the documentation says that it will load data when the page loads. Then, if that is the case, why even use it? I thought the whole point was to only load data when a node is expanded and there are child nodes below it. It also says that manual will only load when user expands a node, so shouldn't it be true or false then? Unless, the documentation is missing something important?

45049

Suggested Answer
[Infragistics] Vince McDonald posted over 15 years ago in reply to chris
intrinsicinnovation said:
But, now I find that there are more than one relationship between some levels of the tree, so I think that manual load on demand would work better?
There's likely a number of options available to represent this kind of multiple relations. Using manual load-on-demand to directly populate the nodes is one of the options. Another is to manipulate your data source so that the underlying data contains all the information from the multiple relations in the same hierarchy.

intrinsicinnovation said:
But, since I already have bound 3 levels and want to add a 4th level, what would be the best way to move this code to manual load on demand?
I would need a clearer idea of what you have already to advise on this. You mention that you already have three separate DataSet objects, which means these have no DataRelation objects linking their tables. This is a useful setup for manual load-on-demand already; you'd simply add a fourth DataSet (or DataTable) along with the logic in the DemandLoad event to read from that new table.

intrinsicinnovation said:
I'm just assuming that manual load on demand would load the tree faster than binding?

Both manual and automatic load-on-demand are quicker than binding (with load-on-demand turned off) since they only creates and renders node objects for data islands that have been displayed.

In general, I expect manual load-on-demand to be quicker than using automatic load-on-demand, since automatic load-on-demand needs to have access to the entire set of data while manual load-on-demand only needs access to the relevant subset of data. Manual load-on-demand also provides additional flexibility for data binding, such as binding to multiple unrelated DataSet objects; conversely, automatic load-on-demand is easier to implement.

intrinsicinnovation said:
Is there some extensive sample code for load on demand manual?
This article from our online Knowledge Base provides additional information about automatic and manual load-on-demand, including a sample implementation of the DemandLoad event for manual load-on-demand:
HOWTO: How to use the LoadOnDemand feature of WebNavigator

As a note, this article was written before we implemented AutomaticSmartCallbacks and ManualSmartCallbacks options for load-on-demand.
Cancel
Reply

Verify Answer

Reject Answer

Cancel
660

chris posted over 15 years ago in reply to [Infragistics] Vince McDonald

Thanks Vince. I already had my tree setup with binding for each level. But, now I find that there are more than one relationship between some levels of the tree, so I think that manual load on demand would work better? But, since I already have bound 3 levels and want to add a 4th level, what would be the best way to move this code to manual load on demand? I'm just assuming that manual load on demand would load the tree faster than binding? I simply call my 'loadtreedata(...)' method when I want to refresh the tree, as the data is always changing in the database, and it has to be refreshed very often, and with about 100 nodes, it's slowing down, as of course it's loading everything at once.

Is there some extensive sample code for load on demand manual? I have 3 datasets, corresponding to 3 levels of the tree.

Thanks
Cancel
Reply

Verify Answer

Cancel
45049

Suggested Answer
[Infragistics] Vince McDonald posted over 15 years ago
Both "Automatic" and "Manual" settings for load-on-demand are triggered in the same way. The root-level nodes are loaded on the initial page load, and when a node is expanded and doesn't yet have any child rows, a postback is triggered to retrieve child rows. They differ in the following ways:

For "Automatic" load-on-demand, you would databind the tree to a hierarchical data source, which contains data appropriate for all levels in the tree. The control takes care of the rest - it retrieves only the root data on initial page load, and determines which subset of child data to retrieve on postbacks triggered to get more data.

For "Manual" load-on-demand, you would either bind the tree to a data source that represents only the first level of data, or you'd manually populate that first level with nodes. You need to handle the DemandLoad event, and in this event provide the logic needed to add the appropriate child nodes to the node that was just expanded.

In short, "Automatic" does the plumbing for you, while "Manual" gives you more control.
Cancel
Reply

Verify Answer

Reject Answer

Cancel