Every now and then, you learn something that seems so obvious that it's completely embarrassing that you didn't know it in the first place. You know, like the first time you realized that it's "intents and purposes", not "intensive purposes," or when you first learned that girls really do make poo. I had a similarly embarrassing moment earlier this week when I learned about XmlParent for the first time.
This one is really my fault for not reading the documentation, but I'm gonna blame it on CFDump anyway. See, like many people I am sure, when I am not positive what's inside of an object or I forget the name of something, I simply dump it out using CFDump. For example, if I wanted to see what was in a ColdFusion XML document, I could run this code:
- <!--- Define the xml document. --->
- <cfxml variable="xmlGirls">
- <name>Christina Cox</name>
- <hair>Dirty Blonde</hair>
- <name>Maria Bello</name>
- <hair>Dirty Blonde</hair>
- <name>Winona Ryder</name>
- <!--- Output ColdFusion XML document. --->
When we run this code, we get the following CFDump output:
| || || || || |
| || |
| || || |
This is great - this really shows us what's in the XML document. Now, this is the "long version" XML output as shown by ColdFusion's CFDump tag, but as it turns out, this is not the complete XML architecture. To avoid what I assume is a circular reference issue, ColdFusion never outputs the XmlParent reference in the CFDump tag. Every child node has an XmlParent reference that allows you to get a reference to the child node's parent, thereby allowing you to walk both up and down the XML tree.
I can't believe I didn't know this!! When I was trying to answer an "Ask Ben" question earlier this week, I literally thought to myself, "There's got to be a better way," and decided, out of the blue, to Google, "XmlParent". Lo and behold, the first search result was the ColdFusion Live Docs' XML entry. How sad is that? I've been working with XML in ColdFusion for how many years? And I just discover XmlParent? That is sooo lame!
So anyway, embarrassment aside, let's just take a look at a quick example. Here, we are going to gather XML nodes via XmlSearch() and then use the resultant nodes' parent element to output data:
- <!--- Gather all hair nodes that are dirty. --->
- <cfset arrDirtyGirls = XmlSearch(
- "//hair[ contains( text(), 'Dirty' ) ]"
- ) />
- Now that we have the hair nodes, loop over them to
- output the parent node (girl) and the hair color.
- Get a reference to the girl. The Girl node is
- the parent of the Hair node and can be accessed
- via XmlParent.
- <cfset xmlGirl = xmlHairNode.XmlParent />
- <!--- Output name and hair color. --->
- #xmlGirl.Name.XmlText# - #xmlGirl.Hair.XmlText#<br />
When we run this code, we get the following output:
Christina Cox - Dirty Blonde
Maria Bello - Dirty Blonde
The example is actually pretty obvious; It's only a shame, it wasn't obvious a few years ago.
Note: I actually have mentioned XmlParent once before on this blog, about a year and a half ago. But, it has long since faded from my memory. I think that's actually MORE embarrassing than having never learned it :)
What? Girls poo? Now I'm crushed. You'll be telling me that dogs fart next, but I refuse to believe it Mr. N. :)
I was as crushed as you are. I tried to fight it for a long time, but it's science.
That girls pooing thing can't be true, because my mom has never been naked in all her life.
You raise a good point. Maybe the whole pooing thing is more of a "pseudo-science". I find that arguments like this cast serious doubts on the whole topic.
Dude, are you spying on me? I was just looking at the results of my xmlSearch thinking how I could get the other data in the element. Totally did not think of the xmlParent since its not in the dump. Thanks.
Word up, they really should put it in the output. Obviously they can't output the values, but they could do something like:
[ XmlParent ][ Output not shown ]
That way, at least we'd remember, oh d'uh, it's in there.