Ask Ben: Working With Inconsistent XML
Posted May 23, 2008 at 10:49 AM
About your post: Testing For The Absence Of A Text Node Using XmlSearch() And XPath... I could not post any code so I am writing you here. Useful info. Here's another one for ya. Say your looping through some XML you got via cfhttp that has, at times, *inconsistent data*. See below.
<girl>
. . . <name>Hayden Panettiere</name>
. . . <age>18</age>
. . . <height></height>
. . . <weight></weight>
. . . <description>
. . . . . . Hayden played Claire, the Cheerleader,
. . . . . . on the hit Fox television show, Heroes.
. . . </description>
</girl>
<girl>
. . . <name>Marisa Miller</name>
. . . <age>26</age>
. . . <description>
. . . . . . Marisa played this girl of my dreams
. . . . . . the last time I slept.
. . . </description>
</girl>
When looping through the girl(s), what do you do when the height or weight nodes simply do not exist like under Marisa Miller? I've used the xmlChildPos() function. I'd like to know what you might use.
XmlChildPos() is a good way to go; I actually only learned about that method when I was researching XML delete functionality. The documentation was really confusing on what it even did. That said, I think one of the tricks to working with XML in ColdFusion is to realize that the ColdFusion XML document is really flexible and diverse as to how it can be accessed and addressed. It really is pretty amazing when you think about it.
For starters, child nodes of a given node can be access as a single array using Node.XmlChildren. Or, they can be accessed as "pseudo" arrays using named addresses such as Node.ChildNode. Furthermore, you can think of these pseudo arrays as belong to a set of sets that is like a pseudo structure. Because of this we can actually access and delete values using ColdFusion array and struct methods. Also, if you want to access the text of a node, you can use Node.XmlText, or you can simply output the node, #Node#. Very easy!
I'm probably not explaining it all well, and I am sure that I am unsure on some of the short-cuts. Actually, that would make a cool blog post in and of itself. That being said, let's take a quick look at how we might handle the inconsistent data above:
Launch code in new window » Download code as text file »
- <!---
- Create XML girls. If this were actually coming from a CFHTTP
- rquest as in the question above, you could just have used
- something like XmlParse( Trim( CFHTTP.FileContent ) ).
- --->
- <cfxml variable="xmlGirls">
-
- <girls>
- <girl>
- <name>Hayden Panettiere</name>
- <age>18</age>
- <height>5'2"</height>
- <weight>125 lbs.</weight>
- <description>
- Hayden played Claire, the Cheerleader, on the
- hit Fox television show, Heroes.
- </description>
- </girl>
- <girl>
- <name>Marisa Miller</name>
- <age>26</age>
- <description>
- Marisa played this girl of my dreams the last
- time I slept.
- </description>
- </girl>
- </girls>
-
- </cfxml>
-
-
- <!--- Loop over girls. --->
- <cfloop
- index="xmlGirl"
- array="#xmlGirls.XmlRoot.XmlChildren#">
-
- Name: #xmlGirl.name#<br />
- Age: #xmlGirl.age#<br />
-
- <!--- Check for height in "pseudo child struct". --->
- <cfif StructKeyExists( xmlGirl, "height" )>
- Height: #xmlGirl.height#<br />
- </cfif>
-
- <!--- Check for weight in "pseudo child struct". --->
- <cfif StructKeyExists( xmlGirl, "weight" )>
- Height: #xmlGirl.weight#<br />
- </cfif>
-
- Description: #xmlGirl.description#<br />
- <br />
-
- </cfloop>
When we run this, we get the following output:
Name: Hayden Panettiere
Age: 18
Height: 5'2"
Height: 125 lbs.
Description: Hayden played Claire, the Cheerleader, on the hit Fox television show, Heroes.
Name: Marisa Miller
Age: 26
Description: Marisa played this girl of my dreams the last time I slept.
Notice that to see if the inconsistent nodes exist, we are merely checking their named existence in the parent node's "pseudo struct set," for lack of a better name. Then, to access all of these values, we are using the name-chain short-hand notation. In fact, we are using several short-hand notations together. Our line of text here:
Launch code in new window » Download code as text file »
- #xmlGirl.name#
... could be rewritten as such:
Launch code in new window » Download code as text file »
- #xmlGirl.name[ 1 ].XmlText#
... which could also be rewritten as such:
Launch code in new window » Download code as text file »
- #xmlGirl.XmlChildren[ 1 ].XmlText#
When you understand the short-hand notation that can be used, you can really find quick ways to check for node existence. Hope that helps.
Download Code Snippet ZIP File
Post Comment | Ask Ben | Other Searches | Print Page
Newer Post
Updated: Converting A ColdFusion Query To CSV Using QueryToCSV()
Older Post
Bug With ARGUMENTS Scope And Implicit Array / Struct Creation
Reader Comments
This is exactly what I do when working with XML. Most of the times I use StructKeyExists(xml, node). We never know when someone is going to touch the xml, delete a node, break the application and make us the guilty ones :)
Ben, using your example, how would you re-write this, a real-world example using the eBay.com api? Would you need to know the complete xml hierarchy in order for this to work? Sorry, I can't post full code directly to your blog...
cfhttp url="someEbayApiUrl" method="GET" result="xmlFeed"
cfset theArray = xmlSearch(xmlFeed, "//*[local-name() = 'Item']")
cfloop index="i" from="1" to="#arrayLen(theArray)#"
cfif xmlChildPos(theArray[i],"PostalCode", 1) GT 0
cfset postalcode = theArray[i].postalcode.xmlText
/cfif
/cfloop
I know this isn't helpful in any way to the code above, but I just wanted to note that Heroes is shown on NBC, not FOX :)
@Gareth,
Ooops :) Good catch.
@Che,
I am not sure what you are asking. I would think that when you work with an API, you need to know the data structure that will be returned... otherwise, how can you possibly know what to do with it?
I used to use
<cfif xmlsearch(xmlgirl,"/height") GT 0>
Height: #xmlGirl.height#<br />
</cfif>
Do you think StructKeyExists( xmlGirl, "height" )> is faster?
@JusufD,
I would assume the StructKeyExists() would be faster, but I am sure both are fast enough.
I'm not familiar with cold fusion. When I am dealing with inconsistent XML I use the XPath:
//cats/item[count > 0]
