_{(Some names have been changed to protect the innocent... ;D)}

Today at work I was asked to create a javascript "class" to manage some basic AJAX requests. By the end of the day I have one working synchronously, but not asynchronously.

My search of the forums revealed suggestions for using JSON (it looks worse than the XML tree) and parsing of XMLHttpRequest.responseText. However, I've been asked to use XML and my class is already designed to parse the XML prettily so those aren't options (and I think I'd prefer to get the XML working anyway).

If you wish to skip the code below and just answer my question I'm trying to create an XMLDOM object from the XMLHttpRequest.responseXML property.

As is, the class is designed similar to a recordset to make it feel like you were simply accessing a database table. For example:

_{The only methods required to implement the class in this example are Request(), RecordCount(), EOF(), and Item().}

When the form loads the above code would draw something like the following:

    RecordCount: 3

    +------------+------------+------------+
    | FieldName1 | FieldName2 | FieldName3 |
    +------------+------------+------------+
    | 1          | foo        | 2352       |
    +------------+------------+------------+
    | 2          | bar        | 34243      |
    +------------+------------+------------+
    | 3          | foo2       | 12         |
    +------------+------------+------------+

Using the following XML:

    <?xml version="1.0" encoding="iso-8859-1"?>
    <Request>
        <Sample>
            <Records>
                
                    1
                    foo
                    2352
                
                
                    2
                    bar
                    34243
                
                
                    3
                    foo2
                    12
                
            </Records>
        </Sample>
        <Sample2>
            <Records>
                
                    The strRootElement is used to select a set of
                                data from potentially more than one "table" of
                                information...
                    
                    foo
                
                
                    bar
                    foo2
                
            </Records>
        </Sample2>
    </Request>

Anybody still reading, thank you. That Request() method of the MyAJAXClass class is currently loading an XML file into an XMLDOM object. As far as I can tell, there is no way to set a callback function for when it finishes loading (asynchronously) so I enforce a synchronous load. The plan, however, is to asynchronously load the xml file and process it when it finishes.

Does anybody know how to define an onLoad callback for the XMLDOM object or preferably how to create an XMLDOM object from the XMLHttpRequest.responseXML?

_{(Some names have been changed to protect the innocent... ;D)}

Let me just say that using XML is crazy dumb. Whoever told you to do that better have an incredibly good reason for doing so. If not, disobey and use JSON.

And regarding kazzmir's code, I've found it's best to check for the ActiveX control first. IE 7's native control is apparently buggy and broken.

I'm mildly impressed myself. The problem isn't with using asynchronous processing. The problem is creating an XML DOM (XMLDOM object) from the XMLHttpRequest object's responseXML property.

Alternatively, if you know how to use the XMLDOM object to load() an XML file asynchronously (as far as I can tell there is no onreadystatechange event for XMLDOM) as you do for XMLHttpRequest that would probably solve my problem as well.

I'm not really a huge fan of XML, but JSON looks pretty ugly to me. Disobeying isn't really an option if I want to keep my job, though I could recommend a change. I don't really see the point. I appreciate the advice, but what I really want/need is to know how to create the XMLDOM object from the XMLHttpRequest, since the processing of XML is already taken care of.

As I mentioned, implementing my class the only method calls I need to access data are the constructor to create the object, Request() to make the request, and from there EOF() and Item() to access the fields.

Noted.

Uhm, no. XML is only better if you need to use XPath to access large amounts of arbitrary data. And even then, browsers' (IE) native XPath support is not good.

If you are dealing with known data (either small datasets or for iteration), JSON is way better.

eval(json);
alert(json.foo);

The only other reason to use XML is if you actually already have well-formed XML documents stored on the server.

That sounds right and I hope it is, but I'm having trouble utilizing it right now. Can you see what might be the problem with this:

It's very possible there is a bug (or many) in my code somewhere, but this is basically blowing up. Firebug detects an uncaught exception (which is messy) and IE7's <select> appears to somewhat "break".

The exception caught by firebug isn't necessarily from the block of code that I showed you... And I'm not working with the code I wrote at work today; I'm using old code I hacked together experimenting with AJAX an unknown number of months ago...

** EDIT **

Actually, after looking up JSON syntax on Wikipedia it actually looks much nicer than the other threads made it seem. I didn't know you could represent objects with JSON... And for the record, it looks a lot nicer when carried over multiple lines.

I still want help parsing the XML from XMLHttpRequest.responseXML, however, I'll also play with JSON and consider recommending it to the company.

Please do...

Its slowly developed over the years. I stopped caring to stop it.

Much appreciated. There's an ASP/VBScript class which might make it a little bit easier to work with. My only fear is licensing issues and whether or not my company would want to use it, though the "license" does say to use it freely so long as the credits are maintained.

** EDIT **

I'm trying to parse the XML and I'm getting an error that gobjXMLDOM.getElementsByTagName("SampleElement")[0] doesn't support the getElementsByTagName method. What am I doing wrong?

I checked with IE's script debugger and there is an object returned by gobjXMLDOM.getElementsByTagName("SampleElement")[0], but it doesn't seem to be the first "SampleElement" element, rather, it seems to be the <?xml version="1.0" encoding="iso-8859-1"?> element...

** EDIT (10 minutes after the first) **

However, when I tried gobjXMLDOM.documentElement.getElementsByTagName("SampleElement")[0] it seemed to work.

Yes, and as fate would have it within minutes of the bump I figured out the problem... Thanks though.