Overview:

This blog post is in response to one of the customer queries we received a couple of days back. One of our India customer called us up and said that they are seeing slowness in couple of pages where they are using RadGrid. Problem statement was – when they click on “Add New Row” button it takes nearly 1.5 minutes for the row to appear on the grid. This made it even scary for them as they had moved the code to production. Through this blog post I will try to outline what I found out and what could have been done better to resolve the problem.

Analysis:

I went in to a call with the customer and took a look at their code. There were couple of things they were doing. Let me list them down:

• DataSource for the RadGrid was a SqlDataSource control defined in the .ASPX page
• They were handling OnItemCommand and OnItemDataBound. They had some logic in this method which would execute on every data bound and every item command
• The new item template had lot of drop down lists and each of them were bound to SqlDataSource control as their data source.

Now, next thing I did was to take look at the timing. We used IE Developer Tools and captured the network traffic when Add New button was clicked. What I found out was, server was taking close to 10 seconds to finish its server side processing. That’s a lot of time if you ask me. What was surprising for them was, they had paging enabled on the grid. So they were thinking that paging is happening correctly but still why is the add new row scenario slow.

So this made me to delve little bit deeper into what exactly is happening in this scenario. So rest of this blog post is all about my observation. Read on if you have faced a similar problem.

Observations:

Did two experiments to figure out what was happening. First I use a RadGrid and bind it to SqlDataSource control and look at the query executed at the DB side. Second was to use entity data source instead of SQL data source and look at the query executed at the DB.

For my experiments I am using Northwind Database and in particular Order Details table. Order Details table contains nearly 2155 records. So this makes a good candidate to mimic a scenario where you are paging the RadGrid lets say 10 records per page.

I assume that you have RadControls for ASP.NET AJAX suite installed on your system. If not, you can always download from our products page – www.telerik.com/products/aspnet-ajax.aspx. You will need that if you want to follow along with this post. Also in order to trace the SQL Queries executed, I am making use of a open source tool called “Express Profiler” which is available at CodePlex here: http://expressprofiler.codeplex.com/. I had to use this as I have a SQL Server 2012 express edition and express edition does not come with Profiler option. You need a full blown SQL server to get the profiler.

I have created a RadControls Web Application for my experiments. When you install the RadControls for ASP.NET AJAX suite, you will get project templates in Visual Studio, using which you can create a web application.

Fig 1: RadControls Web Application Project Template

DataBinding with SqlDataSource control:

First experiment was placing a RadGrid and bind it to SqlDataSource control. I placed a RadGrid and set its basic properties.  Then added a SqlDataSource control to the page and set its connection string and provided a select command. Here is the code snippet for grid and SQL data source control:

 
    <telerik:RadGrid ID="RadGrid1" runat="server" 
                     DataSourceID="SqlDataSource1" 
                     AllowPaging="true" AllowSorting="true" 
                     AllowFilteringByColumn="true" AllowAutomaticInserts="True"
                     MasterTableView-EditMode="InPlace"  
                     MasterTableView-CommandItemDisplay="TopAndBottom" />
    <asp:SqlDataSource ID="SqlDataSource1" runat="server" 
                       ConnectionString="<%$ ConnectionStrings:NorthwindConnectionString %>"
                       ProviderName="System.Data.SqlClient" 
                       SelectCommand="SELECT * FROM [Order Details]" />

Analyzing the queries:

Now, before you run the application, lets get ready to trace the SQL queries that get executed. You should have downloaded the Express Profiler. Start the express profiler – Set the server details, select the authentication mechanism and finally select the events you want to trace. Here is the screen shot:

Fig 2: Express Profiler Settings

Run the web application. Notice that if we just set AllowPaging=true on RadGrid and do not provide a page size, by default RadGrid puts 10 as the PageSize. So in my case I should see only 10 rows on my first page of the RadGrid. Here is the Grid output:

Fig 3: RadGrid output

When I ran the app and looked at the trace, here is what I get:

Fig 4: SQL Query Trace for SQL Data Source Control bound to RadGrid

What you see here is a call to retrieve all the records from the table. You may be asking but the page size is set to 10. That is correct but what really happens is this: RadGrid will go ahead and ask SQL Data Source to get the data. SQL Data Source will blindly execute the query that has been set for the select command. RadGrid will not be able to do any tinkering with query because the SQL Data Source does not provide us a chance to put paging related statements.

Now lets do a page change on the grid. I will go ahead and click on Page 2.  RadGrid will now navigate to Page 2. Here is the trace that is generated for this:

Fig 5: SQL Trace for Page Change Event

What you notice is, the SQL Data Source goes ahead and brings all the records from the table. RadGrid will then pick up only the second page rows and displays it.

Now lets go ahead and click on the Add New Row button and trace the query executed:

Fig 6: SQL Trace for Add New Record Event

As you can see it again retrieves all the records from the table.

What this tells us is the time taken is due to the fact that its going and getting all the table records no matter whatever is the situation we have. Now hitting DB and getting everything does call for some time although small but it till takes up time on the server.

Now lets do the same experiment but with Entity Data Source.

DataBinding with Entity Data Source:

I will use the same data grid but instead of SQL Data Source I will use a Entity Data Source. In order to work with Entity Data Source, you need a Entity Data Model. I will not get into how to create a Entity Data Model but assume that you know how to do it. Just create a ADO.NET Entity Data Model and generate the models from the database and use the same Northwind database. Once you have the ADO.NET Entity Data Model created, create a Entity Data Source on the page. Here is the Entity Data Source code snippet that I have:

 
<asp:EntityDataSource ID="EntityDataSource1" runat="server" 
                          OnContextCreating="EDS_ContextCreating" 
                          EntitySetName="Order_Details" />

I am making use of the OnContextCreating event to set the DbContext of ADO.NET Data Model to Entity Data Source ObjectContext. Here is the code for the event handler of OnContextCreating:

 
protected void EDS_ContextCreating(object sender, EntityDataSourceContextCreatingEventArgs e)
{
    var db = new NORTHWINDEntities();
    e.Context = (db as IObjectContextAdapter).ObjectContext;
}

Here NORTHWINDEntities is my Entity Data Model context. I am just casting it to ObjectContext since Entity Data Source works with ObjectContext. I set this entity data source as the data source of the RadGrid.

Analyzing the Queries:

Now lets run the app and trace SQL queries for the same set of actions we did with SQL Data Source control.

First, on page load:

Fig 7: Entity Data Source – Page Load – Query Trace

As you can see we have put in a “SELECT TOP 10” statement in the query. So its not bringing all 2155 records rather only 10 records. So time spent to do this will be very negligible. We also execute one additional query to get the total record count.

Now lets see what happens when we do a paging:

Fig 8: Entity Data Source – Paging – Query Trace

Now, lets trace the query when we do ‘’Add new record” scenario:

Fig 9: Entity Data Source – Add New Record – Query Trace

Notice that, in this scenario, we automatically page to the last row and add a blank record.

As you can see from the above experiments, RadGrid is intelligent enough to optimize the queries based on the scenario you are dealing with.

Conclusion:

It was very evident from the experiments that if you are using SQL Data Source as your data source choice, there is not much RadGrid can do to optimize the query before it gets executed. Where as with Entity Data Source, RadGrid is smart enough to optimize the query and always have a optimized data fetch. I personally recommend using an ORM in the project for the data access scenarios and use the Entity Data Source as your data source choice. We i.e at Telerik have a free ORM called OpenAccess. Its free for life so you can pretty much use this in your production without any hassle at all. Do give it a try and let us know how you feel about it.

I hope this blog post does give you information on things you need to take care while doing data binding to Grids. This is just not in particular to RadGrid, this is mostly about any data binding scenario. Trace your queries and be sure that write and optimized queries are getting executed in your apps.

Till next time, Happy Coding.