“~>” Notation

The second non-standard syntax is to provide a convenient way to extract values from multiple elements in the XML. In the following example, we are attempting to index the entry and want to extract the full name of every contact:

<root>
  <entry>
    <contact>
      <firstname>Fred</firstname>
      <lastname>Bloggs</lastname>
    </contact>
    <contact>
      <firstname>John</firstname>
      <lastname>Smith</lastname>
    </contact>
  </entry>
</root>

Our implementation allows the following XPath expression to concatenate values from multiple elements:

    "concat(~/contact/firstname, ' ', ~/contact/lastname)"

Unfortunately, this actually concatenates the first values from each inner expression and returns only the string

    “Fred Bloggs”

The following expression allows all of the full names to be returned (this can be used when constructing metadata, or can be used with the xml:XPathValues function with Lua):

    "~/contact ~> concat(~/firstname, ' ', ~/lastname)"

When used with the xml:XPathValues function in Lua, as below,

    xml:XPathValues("~/contact ~> concat(~/firstname, ' ', ~/lastname)")

this returns the following Lua value:

    { "Fred Bloggs", "John Smith" }

The new syntax added here is the operator “~>”.

    "LeftXPath ~> RightXPath"

The value on the left of the "~>" is parsed as an XPath expression, and should return a set of elements.

The value on the right of the "~>" is parsed as an XPath expression, and is processed relative to all the elements returned by the left side.

For more information on the normal XPath syntax supported see http://en.wikipedia.org/wiki/XPath_1.0.


_HP_HTML5_bannerTitle.htm