RE: [flexcoders] E4X in Flex 2.0, part 1: Reading XML

Tue, 17 May 2005 17:04:49 -0700

Title: Message

This is very exciting!  I’m especially excited about the double-dot and wildcard operators.  Thanks for this update, as it’ll help with planning for future projects.  Keep ‘em coming!

 

 

/******************************************
* Jeff Beeman
******************************************/

From: flexcoders@yahoogroups.com [mailto:[EMAIL PROTECTED] On Behalf Of Gordon Smith
Sent: Tuesday, May 17, 2005 4:54 PM
To: 'flexcoders@yahoogroups.com'
Subject: [flexcoders] E4X in Flex 2.0, part 1: Reading XML

 

As you may have gathered, we've been spending a lot of time lately leveraging the new features of the Flash Player in the new Flex application model. Naturally, you'll also be able to leverage those same new features, so we thought we'd start giving you a run down of what's new. Of course we don't have beta software for you to play with yet, so for now, we'll provide a lot of detail so you can evaluate these new features and give guidance for us.

 

XML manipulation in Flex 2.0 is going to get a lot more powerful, as well as faster. By the time that Flex 2.0 ships, the Flash Player will support E4X ("ECMAScript for XML"), a set of programming language extensions adding native XML support to ECMAScript. The player team is busy implementing Standard ECMA-357 as described in http://www.ecma-international.org/publications/standards/Ecma-357.htm.

 

Here's how the spec describes what this feature offers: "E4X adds native XML datatypes to the ECMAScript language, extends the semantics of familiar ECMAScript operators for manipulating XML objects and adds a small set of new operators for common XML operations, such as searching and filtering. It also adds support for XML literals, namespaces, qualified names and other mechanisms to facilitate XML processing."

 

Lets take a look at a few examples of how you can read XML data using E4X.

 

As in the current player, you'll be able to create variables of type XML by parsing a String. But XML literals will now be supported as well:

 

    var employees:XML =
        <employees>
            <employee ssn="123-123-1234">
                <name first="John" last="Doe"/>
                <address>
                    <street>11 Main St.</street>
                    <city>San Francisco</city>
                    <state>CA</state>
                    <zip>98765</zip>
                </address>
            </employee>
            <employee ssn="789-789-7890">
                <name first="Mary" last="Roe"/>
                <address>
                    <street>99 Broad St.</street>
                    <city>Newton</city>
                    <state>MA</state>
                    <zip>01234</zip>
                </address>
            </employee>
        </employees>;

 

Instead of using DOM-style APIs like firstChild, nextSibling, etc., with E4X you just "dot down" to grab the node you want. Multiple nodes are indexable with [n], similar to the elements of an Array:

 

    trace(employees.employee[0].address.zip);

    ---

    98765

 

To grab an attribute, you just use the .@ operator:

 

    trace([EMAIL PROTECTED]);
    ---

    789-789-7890

 

If you don't pick out a particular node, you get all of them, as an indexable list:

 

    trace(employees.employee.name);

    ---

    <name first="John" last="Doe"/>

    <name first="Mary" last="Roe"/>

 

(And note that nodes even toString() themselves into formatted XML!)

 

A handy double-dot operator lets you omit the "path" down into the XML _expression_, so you could shorten the previous three examples to

 

    trace(employees..zip[0]);

    trace([EMAIL PROTECTED]]);

    trace(employees..name);

 

You can use a * wildcard to get a list of multiple nodes or attributes with various names, and the resulting list is indexable:

 

    trace(employees.employee[0].address.*);

    ---

    <street>11 Main St.</street>

    <city>San Francisco</city>

    <state>CA</state>

    <zip>98765</zip>

    trace([EMAIL PROTECTED]);
    ---

    Doe

 

You don't have to hard-code the identifiers for the nodes or attributes... they can themselves be variables:

 

    var whichNode:String = "zip";
    trace(employees.employee[0].address[whichNode]);

    ---

    98765

 

    var whichAttribute:String = "ssn";
    trace([EMAIL PROTECTED]);
    ---

    789-789-7890

 

A new for-each loop lets you loop over multiple nodes or attributes:

 

    for each (var ssn:XML in [EMAIL PROTECTED])

    {

        trace(ssn);

    }

    ---

    123-123-1234
    789-789-7890

 

Most powerful of all, E4X supports "predicate filtering" using the syntax .(condition), which lets you pick out nodes or attributes that meet a condition you specify using a Boolean _expression_. For example, you can pick out the employee with a particular social security number like this, and get her state:

 

    var ssnToFind:String = "789-789-7890";
    trace(employees.employee.(@ssn == ssnToFind)..state);
    ---

    MA

 

Instead of using a simple conditional operator like ==, you can also write a complicated predicate filtering function to pick out the data you need.

 

By the way, although none of my examples use XML namespaces, E4X has complete support for them.

 

Compared with the current XML support in the Flash Player, E4X allows you to write less code and execute it faster because more processing can be done at the native speed of C++.

 

Since E4X is so powerful, we're working to make Flex 2.0 play nicely with it. Components like List, ComboBox, and DataGrid will be able to accept E4X expressions like employees..name as a dataProvider. The <mx:XML> tag will be able to declare an E4X-style XML object in MXML. WebService and HTTPService will be able to deliver E4X-style XML objects across the wire, and they'll use the speed of E4X to do their own work (such as constructing and desconstrucing SOAP packets) faster.

 

For backward compability, the new player will continue to support the old-style XML and XMLNode objects, with the one change that the old XML class will be renamed to XMLDocument. The renaming is necessary because the E4X standard specifies that the new type be called XML. So if you have XML code and you want to continue using, just use XMLDocument instead. But we'll be encouraging everyone to migrate to E4X because it is so much simpler, faster, and more expressive.

 

Please let us know whether you'll find E4X useful in your particular applications.

 

- Gordon


 


 

 

 


Yahoo! Groups Links

Reply via email to