Here is the XML data that will be used in this example.

<addressbook>
  <address firstname="Clark" lastname="Kent">
    <street>344 Clinton St., Apt. 3B</street>
    <city>Metropolis</city>
  </address>

  <address firstname="Sherlock" lastname="Holmes">
    <street>221B Baker Street</street>
    <city>London</city>
  </address>

  <address firstname="Homer" lastname="Simpson">
    <street>742 Evergreen Terrace</street>
    <city>Springfield</city>
  </address>
</addressbook>

node[] main(document in) {
}

This demonstrates another possible main function signature. The code declares main to output a stream of XML data and read standard input as an XML document.

Next we select and output the address with firstname attribute equal to "Sherlock" for output.

node[] main(document in) {
  in/addressbook/address[@firstname == "Sherlock"];
}

XPath programmers will notice a difference here. XmlPL, like C used '==' for equality and '=' for assignment unlike XPath where '=' is the equality operator. Also, notice that in is a variable reference and not a path step such as addressbook and address.

The following comands will build and run the program.

$ xmlplcc query.xml
$ ./query < addressbook.xml

This should be the output.

<?xml version="1.0" encoding="UTF-8"?>
<address firstname="Sherlock" lastname="Holmes">
    <street>221B Baker Street</street>
    <city>London</city>
  </address>

The odd indenting of the closing </address> tag is due to whitespace in the input file. You can clean up the output with a program such as tidy using the following command.

$ ./query < addressbook.xml | tidy -xml