BizTalk XSLT & XPath Examples – Substring and Nested XPath with Where Clause

March 31, 2016
Leave a comment

Two handy references for BizTalk XSLT:

Substring in XSLT Value-Of XPATH

The first example is fairly easy, I just needed to get the date/time without the timezone modifier, in other words to convert this: 2016-01-29T00:00:00-08:00

to this: 2016-01-29T00:00:00.

BizTalk_XSLT_XPath_Substring
The reason was that our web service was only coded to handle the date without the timezone. It used an Orcacle date/time conversion like this:

eValue = "TO_DATE('" + eValue.Replace("T", " ").Replace(".0000000", "") + "','YYYY-MM-DD HH24:MI:SS')";

XPath “Where” clause when using complex namespaces

Take a look at the data below. I needed to map the EventDate associated with DateType=PLN to D02, and the EventDate associated with DataType=FCT to D03.

Biztalk_Xpath_XSLT_Sample_Data

Most of the examples on the internet show you how to do this when you don’t have namespaces involved. But when you do have namespaces, the trick is to use nested queries in your XPATH.  In the sample below, you can see that the text in yellow is complete inside the other xpath brackets.

Using [local-name()=’xxx”] is like saying find me the element whose name matches ‘xxx’.

Once the where has been matched, the “xsl:value-of” statement begins with ./ to go up one level, then to find the EventDate (that is on the same level as DateType) to put in the output field.

BizTalk_XSLT_Nested_Xpath_Where_Clause

Online XPath Tester

December 2, 2015
Leave a comment

XpatherTester.com is the current “go to” tool I use when I want to test XPath (online and for free). I prefer the one built-in to Stylus Studio, but that’s a paid for product that I cannot install on each client’s computer.

To actually learn XPath, I suggest the W3Schools site: http://www.w3schools.com/xsl/xpath_intro.asp

To test the tool, you can copy/paste a sample XML from here: https://msdn.microsoft.com/en-us/library/ms762271%28v=vs.85%29.aspx

Step 1 – Paste in your XML to the tool

XPathTester_Step1_Paste_Your_XML

Step 2- Type in your XPath

For example, I typed in //book/title below. I have highlighted in yellow, because at first, it’s not intuitively obvious where to actually type in the XPath. Note: Do not type in the XPath in the Namespaces: box!

XPathTester_Step2_Type_Your_XPath

Step 3- View your results

The tool will leave your original XML on one tab of your browser, then the results will be opened in a new tab on your browser. This picture shows the original tab on the left, and the new tab on the right.

XPathTester_broiwser_tabs
Note that the “” xml wrapper is added to keep the results as well-formed XML. When you run your XPath in C# or other tools, that will not happen.

XPathTester_Step3_Results

Step 4- Retry as needed

When doing a complicated XPath, you might have to try several times. Close the results tab in your browser, and go back to your original tab to try again.

If you want to specify something like //ns1:books you can define namespace ns1 in the namespaces textarea at the top of the screen.  I haven’t placed with that much.  I tend to create XPath that uses the “localname” syntax, which is a little more difficult.  Use of this syntax is explained well on this StackOverflow page: http://stackoverflow.com/questions/10813653/xpath-select-node-based-in-a-condition-with-local-name.

For example:

//*[local-name()='a'][*[local-name()='aCode']='aaa']

Additional XPATH resources/samples for dealing with local-name and namespaces

Here are some old posts of mine when learning to do XPath back in 2004, it hasn’t really changed since then:

  1. http://www.stylusstudio.com/SSDN/default.asp?action=9&fid=23&read=3074
  2. http://www.stylusstudio.com/SSDN/default.asp?action=9&read=5018&fid=48&FirstTopic=1420&LastTopic=1439&style=WebBoard

Xpath – selecting element by attribute value

March 10, 2015
Leave a comment

Using the sample data below, I wanted to get all the Link node that have LinkID = 13 (not the Junk node).

Sample data

<whatever>
  <Links>
      <Link LinkID='13' /> 
      <Junk LinkID='13' />
      <Link LinkID='1' /> 
      <Link LinkID='13' /> 
  </Links>
</whatever>

What fails are these:

//*[local-name()=’Link’]/*[@LinkID=’13’] (XPath returned 0 items)
//*[local-name()=’Links’]/[local-name()=’Links’ and @LinkID=’13’] ( Unexpected token “[” in path expression or Expression must evaluate to a node-set)
//[local-name()=’Links’]//*[local-name()=’Link’][@LinkID=’13’] ( Unexpected token “[” in path expression or Expression must evaluate to a node-set) – This is same as one that worked below except // instead of //*

What works are these:

//*[local-name()=’Link’ and @LinkID=’13’]
//*[local-name()=’Link’][@LinkID=’13’]
//*[local-name()=’Links’]//*[local-name()=’Link’][@LinkID=’13’]

Complex XPath

February 2, 2009
Leave a comment

Complex XPath with BizTalk Schema/Data

http://www.w3schools.com/Xpath/xpath_examples.asp does a pretty good job of showing elementary XPath.
But in the “Real World” with BizTalk, Xpath can be more challenging.
This is a follow-up to my January 3, 2005 blog entitled “Common XPATH Examples, Questions, Issues with Biztalk”
http://nealwalters.blogspot.com/2005/01/common-xpath-examples-questions-issues.html

When BizTalk creates a sample instance, it creates it in this structure (except below, more realistic data has been entered).

<books>
   <book>
     <title>A Tale of Two Cities</title>
     <authorLastname>Dickens</authorLastname>
     <authorFirstname>Charles</authorFirstname>
     <category>fiction</category>
   </book>
   <book>
     <title>Operating Manual for Spaceship Earth</title>
     <authorLastname>Fuller</authorLastname>
     <authorFirstname>Buckminster</authorFirstname>
     <category>science</category>
   </book>
   <book>
     <title>The Singularity Is Near: When Humans Transcend Biology</title>
     <authorLastname>Kurzweil </authorLastname>
     <authorFirstname>Ray</authorFirstname>
     <category>science</category>
   </book>
   <book>
     <title>The Age of Spiritual Machines: When Computers Exceed Human Intelligence</title>
     <authorLastname>Kurzweil </authorLastname>
     <authorFirstname>Ray</authorFirstname>
     <category>science</category>
   </book>
</books>

/books/book/title – returns all titles

/books/book/category[.=’science’] – returns all categories, where the text of that category = ‘science’

/books/book[category=’science’]/title – returns all titles, where the text of that category = ‘science’

/books/book[category=’science’ and authorFirstname=’Ray’]/title – return all titles where Ray is the author and category = ‘science’

When BizTalk creates a sample instance, it creates it in this structure (except below, more realistic data has been entered).
Each “book” record has it’ s own name space. This data was created with a ‘books’ schema, which in turn imported a “book” schema.
The book record was defined by adding a new “record”, and settings its property “Data Structure (Type):

<ns0:books xmlns:ns0=”http://biztalk-training.com/books”>
   <ns1:book xmlns:ns1=”http://biztalk-training.com/book”>
     <title>A Tale of Two Cities</title>
     <authorLastname>Dickens</authorLastname>
     <authorFirstname>Charles</authorFirstname>
     <category>fiction</category>
   </ns1:book>
   <ns1:book xmlns:ns1=”http://biztalk-training.com/book”>
     <title>Operating Manual for Spaceship Earth</title>
     <authorLastname>Fuller</authorLastname>
     <authorFirstname>Buckminster</authorFirstname>
     <category>science</category>
   </ns1:book>
   <ns1:book xmlns:ns1=”http://biztalk-training.com/book”>
     <title>The Singularity Is Near: When Humans Transcend Biology</title>
     <authorLastname>Kurzweil</authorLastname>
     <authorFirstname>Ray</authorFirstname>
     <category>science</category>
   </ns1:book>
   <ns1:book xmlns:ns1=”http://biztalk-training.com/book”>
     <title>The Age of Spiritual Machines: When Computers Exceed Human Intelligence</title>
     <authorLastname>Kurzweil</authorLastname>
     <authorFirstname>Ray</authorFirstname>
     <category>science</category>
   </ns1:book>
</ns0:books>

When BizTalk creates a file from the mapping utility, it combines the namespaces onto the root element, and uses a different prefix, as shown below:

<ns0:books xmlns:ns0=”http://biztalk-training.com/books” xmlns:bk=”http://biztalk-training.com/book”>
   <bk:book>
     <title>A Tale of Two Cities</title>
     <authorLastname>Dickens</authorLastname>
     <authorFirstname>Charles</authorFirstname>
     <category>fiction</category>
   </bk:book>
   <bk:book>
     <title>Operating Manual for Spaceship Earth</title>
     <authorLastname>Fuller</authorLastname>
     <authorFirstname>Buckminster</authorFirstname>
     <category>science</category>
   </bk:book>
   <bk:book>
     <title>The Singularity Is Near: When Humans Transcend Biology</title>
     <authorLastname>Kurzweil</authorLastname>
     <authorFirstname>Ray</authorFirstname>
     <category>science</category>
   </bk:book>
   <bk:book>
     <title>The Age of Spiritual Machines: When Computers Exceed Human Intelligence</title>
     <authorLastname>Kurzweil</authorLastname>
     <authorFirstname>Ray</authorFirstname>
     <category>science</category>
   </bk:book>
</ns0:books>

/*[local-name()=’books’ and namespace-uri()=’http://biztalk-training.com/books’]/*[local-name()=’book’ and namespace-uri()=’http://biztalk-training.com/book’]/title

/*[local-name()=’books’]/*[local-name()=’book’]/title

/*[local-name()=’books’]/*[local-name()=’book’][category[.=’science’]]/title

If for some reason, you had a namespace on category, the same query could be done like this:
/*[local-name()=’books’]/*[local-name()=’book’][*[local-name()=’category’][.=’science’]]/title

/*[local-name()=’books’]/*[local-name()=’book’][category=’science’ and authorFirstname=’Ray’]/title

Thanks to the support team at “StylusStudio.com” for helping with these concepts in their forums. Click here for –>
My discussion thread on these issues
The trick they revealed is that the [] expressions can be nested. Here’s another post, where we talked about selecting a higher node:
link.

That’s the end of the XPath lesson, if you want to see how the “Books” schema was created, see here… BizTalk Schema Imports.