Comments on: Deleting XML nodes: harder than it looks

By: Vic Mortelmans

Vic Mortelmans — Fri, 18 Dec 2009 21:24:18 +0000

So if I want to delete all elements that match some xpath, this is the only (clumsy) way to go?:

var nrofitemstodelete = queryoutput..div.(@["class"] == ‘opomba-layer’).length();
for (i=0;i delete queryoutput..div.(@["class"] == 'opomba-layer')[0];
}

Seems like a performance penalty, since the xpath is evaluated over and over.

Can one evaluate the delete, to see if it actually deleted an item? Then a while-loop would be more appropriate.

Groeten,
Vic

By: Liza

Liza — Fri, 24 Apr 2009 11:19:37 +0000

If you ever want to hear a reader’s feedback :) , I rate this post for four from five. Decent info, but I just have to go to that damn msn to find the missed bits. Thanks, anyway!

By: brio1337

brio1337 — Tue, 24 Feb 2009 22:37:43 +0000

kanenas, I agree, that’s the exact same piece of code I just determined for deleting a known XML node from its parent.

quinthar, thanks for your article, I agree that deleting XML nodes is harder than it looks! Part of the reason is that Adobe doesn’t explain very well how these XML and XMLList objects are related. I see some wording in your article that indicates a misunderstanding.
When you discuss child.appendChild(grandChild), you say that you expected it to return a reference to the node it just “allocated”. But there is no node allocated: grandChild is changed to be part of child’s tree. So now grandChild’s parent() and childIndex() are now valid. This has some consequences, such as you can’t take a node and append it to several other XML nodes and expect it to be in all of them. (You can use the XML object’s copy() function for that).

Similarly, when you call children() on an XML object, the XMLList returned is a reference to the actual child list inside the XML object. So if you delete a node from this list, you’ve actually deleted it from the parent!

I wish Adobe was clearer about this stuff…

By: kanenas

kanenas — Thu, 22 Jan 2009 14:04:16 +0000

The delete operator is indeed the official way (as mentioned in the "appendChild()" documentation) to remove a child node. Perhaps the "Attempt to delete the fixed property..." error arises because grandChild is not the result of an E4X expression? Whatever the cause, you can get a child node's index with the childIndex() method: delete grandChild.parent().children()[grandChild.childIndex()] A bit lengthy, .

By: noah a-b

noah a-b — Mon, 24 Nov 2008 21:55:21 +0000

Wow! They can make things so darn tricky it’s, like searching for a needle in a haystack, only the needle could also randomly be disguised as either a thumbtack, a bobby pin, a screw, or a paperclip!

Your post helped me a lot, but I must say it was this little nugget that finally solved my node deletion issue!

“… you have to use the [0] index or the delete function will puke.”

Praise the code demons!

By: Jeremy

Jeremy — Tue, 21 Oct 2008 02:26:53 +0000

This post didn’t solve my problem directly, but you got me headed in the right direction so I thought I’d share the solution.

I needed to remove an arbitrary node at an arbitrary depth in an XML object. The solution is in two parts.

First, each element that will be deleted needs a UID so I used the UIDUtil.createUID(); function.

var xDoc:XML = XML(”);

I get the item to delete from a Tree.selectedItem call

var delItem:XML = XML(Tree.selectedItem);

Then the delete command is:

delete xDoc.descendants().(@uid==delItem.@uid)[0];

That searches recursively for all elements that have a uid parameter and matches them to the selected item’s uid. Since the return value is an XMLList (even if only 1 item is returned) you have to use the [0] index or the delete function will puke.

Enjoy :)