ASP.NET, C# and Anonymous Types - Iterate through a DataTable while Manually Building an Anonymous Type

Question

Dear Group,

I am currently implementing a client-side paging solution using ASP.NET, jQuery and JSON.

I have been following the excellent article from encosia: http://encosia.com/2008/08/20/easily-build-powerful-client-side-ajax-paging-using-jquery/

In my Web Method I retrieve my data from the database as a DataTable:

    DataTable categoryProducts = ProductViewerAccess.GetCategoryProducts
        ("AA", 4, 0, Page.ToString(), out howManyPages, "FALSE", 0, "CostPrice", "asc", destinationList);

I then retrieve the data from the DataTable into an anonymous type:

    var feeds =
        from feed in categoryProducts.AsEnumerable()
        select new
        {
            Description = feed.Field<string>("description"),
            MfPartNo = feed.Field<string>("MfPN"),
            Inventory = feed.Field<Int32>("Inventory")
        };

The anonymous type is then returned from the Web Method to the client-side:

return feeds.Take(PageSize);

A template then extracts and displays the fields:

  <tbody>
    {#foreach $T.d as post}
    <tr>
      <td>
        {$T.post.Description}
        <p>Mfr#: {$T.post.MfPartNo}</p>
      </td>
      <td>{$T.post.Inventory}</td>
    </tr>
    {#/for}
  </tbody>

This all works great.

However, I would like to extend the code to perform some evaluation checks (e.g., check that various columns in the DataTable are not NULL) and other pre-processing (e.g., call various functions to build the image URL based on the image ID - which is another column in the DataTable not shown in the code fragment) before I return the resulting rows of the DataTable as an anonymous type to the client-side.

Basically, I want to iterate through the DataTable, perform the evaluation checks and pre-processing, while building my anonymous type manually as I go. Or maybe there is a better way to achieve this?

Is there anyway I can achieve this?

Kind Regards

Walter

Answer 1

Sometimes I find it helpful to use the "let" keyword using LINQ to store value I need to use later in the query. I can use this variable later for simple null checks or other things. For example:

var feeds =
    from feed in categoryProducts.AsEnumerable()
    let MfPN = feed.Field<string>("MfPN")
    // Get image url if MfPN is not null, else return default image url.
    let Url = !string.IsNullOrEmpty(MfPN) ? GetImgUrl(MfPN) : “DefaultImage.jpg”
        select new
        {
            Description = feed.Field<string>("description"),
            MfPartNo = MfPN,
            Inventory = feed.Field<Int32>("Inventory"),
            ImageUrl = Url
        };

The only other thing I can think if is too simple perform your pre-processing on the DataTable before calling the LINQ Query.

Hope this helps.

Answer 2

I think that checking for nulls is probably something that makes sense on the server-side. The approach that Douglas describes is one workable one. Another is to handle the null issue as you're building the collection of anonymous types:

var feeds =
    from feed in categoryProducts.AsEnumerable()
    select new
    {
        Description = feed.Field<string>("description"),
        MfPartNo = feed.Field<string>("MfPN"),
        // Return 0 if the inventory value is null.
        Inventory = (int?)feed.Field("Inventory") ?? 0
    };

ScottGu has a good post on using the null coalescing operator to handle nulls concisely, as shown above.

As for building the links, that's something I would probably suggest doing in the client-side template. You can eliminate quite a bit of redundant data sent in the JSON that way. Something like this, for example:

<tbody>
  {#foreach $T.d as post}
  <tr>
    <td>
      <a href="/url/to/details.aspx?id={$T.post.ID}">{$T.post.Description}</a>
      <p>Mfr#: {$T.post.MfPartNo}</p>
    </td>
    <td>{$T.post.Inventory}</td>
  </tr>
  {#/for}
</tbody>