[whatwg] getElementsByIdPart

[chriswa at comcast.net]

Ok, I finally worked up an example using XPATH that does what I want.

document.getElementsById = function(term)
 {
    var xpathString = "//*[contains(@id,'" + term.toString() + "')]"
    var xpathResult = document.evaluate(xpathString, document, null, 0, null);
    var outArray = new Array();
    while ((outArray[outArray.length] = xpathResult.iterateNext())) { }
    return outArray;
 }

 function OnLoad()
 {
    //alert( "loading" );
    debugger; var elements = document.getElementsById( "FormItem1" );
    alert( elements.length ); alert( elements.length );
    for ( var ii = 0; ii < elements.length; ++ii )
    {
        alert( ii + ": ==> " + elements[ii].id );
    }
} 

This works in Mozilla and I can use the indexOf technique for IE.

Thanks to everyone who challenged my assumptions.

Chris


> Thanks Ian for your response.
> 
> What I'm looking for is a way to apply a regular expression that can return 
> matched elements.  What I am doing is taking advantage of the way ASP.NET 
> generates id names and use javascript to group them together as structures.
> 
> For example:
> 
> a - <input id="outer1_inner1_element">
> b - <input id="outer1_inner1_image">
> c - <input id="outer1_inner2_element">
> d - <input id="outer1_inner2_image">
> e - <input id="outer2_inner1_image">
> f - <input id="outer2_inner1_image">
> 
> What I'd like to do is the following:
> 
> var elOuter1 = getElementsById(/outer1/)   // returns elements a - d
> var elOuter1_Inner1 = getElementsById(/outer1_inner1/)  //  returns elements a, 
> b
> 
> The point is that rather than having an exact match as getElementsByName does 
> now you can match on segments of the id attribute.  Currently the only way to do 
> this now is to do the following:
> 
> var inputElements = getElementsByTagName( "input" );
> var elOuter1 = [];
> for ( var ii = 0; ii < inputElements.length; ++ii )
> {
>     if ( inputElements[ii].indexOf( "outer1" )  != -1 )
>          elOuter[ii] = inputElments[ii];
> }
> 
> I am assuming that having an instrinic function (most likely written in C++) 
> would be more performant than doing the above JavaScript and since the 
> information is already available to the DOM and I'm assuming that this would be 
> a trival function to implement.  I appreciate the suggessions to use XPATH but I 
> haven't found an easy, straightforward and compatible solution (as HTML docs are 
> incompatable).
> 
> Thanks,
> Chris
> 
>  
> 
> 
> 
> > On Sat, 21 Aug 2004, Doron Rosenberg wrote:
> > >
> > > Wouldn't a getElementsByAttribute be a more flexiable solution?
> > 
> > It depends, what exactly did you have in mind?
> > 
> > Note that the "." selector in CSS can't be directly emulated even with a 
> > single "[ =~ ]" selector.
> > 
> > I think for most uses, DOM Traversal is the way to go here. 
> > getElementsByClass is IMHO an exception, in that implementing it as a node 
> > filter is non-trivial.
> > 
> > -- 
> > Ian Hickson               U+1047E                )\._.,--....,'``.    fL
> > http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
> > Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'