Read values from an XML file

Goal: Pull elements, text, and attributes out of an XML file such as a Maven pom.xml or an RSS feed, and print them as a Markdown table or list.

Prerequisites: The built-in xml module. Read the file with -I xml, which parses it into a tree of {tag, attributes, children, text} dicts. The example also uses the built-in csv module to render a table.

Query

The dependencies of a pom.xml as a Markdown table:

$ mq -I xml 'import "xml" | import "csv" | xml::xml_find_all("dependency") | map(fn(d): {"group": xml::xml_text(xml::xml_find(d, "groupId")), "artifact": xml::xml_text(xml::xml_find(d, "artifactId")), "version": xml::xml_text(xml::xml_find(d, "version"))};) | csv::csv_to_markdown_table()' pom.xml

Input (pom.xml)

<?xml version="1.0" encoding="UTF-8"?>
<project>
  <modelVersion>4.0.0</modelVersion>
  <artifactId>demo-app</artifactId>
  <version>1.2.0</version>
  <dependencies>
    <dependency>
      <groupId>org.slf4j</groupId>
      <artifactId>slf4j-api</artifactId>
      <version>2.0.9</version>
    </dependency>
    <dependency>
      <groupId>com.google.guava</groupId>
      <artifactId>guava</artifactId>
      <version>33.0.0-jre</version>
      <scope>test</scope>
    </dependency>
  </dependencies>
</project>

Output

| group | artifact | version |
| --- | --- | --- |
| org.slf4j | slf4j-api | 2.0.9 |
| com.google.guava | guava | 33.0.0-jre |

Filter by an attribute

xml_attr(element, name) returns an attribute’s value, or None when it is not set. This lists the id of every <user> whose role is dev:

$ echo '<users><user id="1" role="admin"/><user id="2" role="dev"/><user id="3" role="dev"/></users>' | mq -I xml 'import "xml" | xml::xml_find_all("user") | filter(fn(u): xml::xml_attr(u, "role") == "dev";) | map(fn(u): xml::xml_attr(u, "id");)'

["2", "3"]

Turn an RSS feed into a list

$ curl -s https://example.com/feed.xml | mq -I xml 'import "xml" | xml::xml_find_all("item") | map(fn(i): "- [" + xml::xml_text(xml::xml_find(i, "title")) + "](" + xml::xml_text(xml::xml_find(i, "link")) + ")";) | join("\n")'

With a feed that has two <item> entries:

- [v0.8.5](https://example.com/v0.8.5)
- [v0.8.4](https://example.com/v0.8.4)

Notes

• xml_find_all(tree, tag) returns every element named tag in document order, including the root itself. xml_find returns only the first match, or None when nothing matches. Because matching is at any depth, xml_find(tree, "version") on the pom.xml above returns the project’s own 1.2.0, since it comes before the dependency versions.
• xml_text(element) joins the text of an element and all its descendants with single spaces. For <user><name>Alice</name><email>[email protected]</email></user> it returns Alice [email protected].
• Keep only some elements by testing a child: filter(fn(d): xml::xml_text(xml::xml_find(d, "scope")) == "test";) keeps the test-scoped dependencies of the pom.xml above. Elements without a <scope> do not match.
• -I xml is shorthand for reading with -I raw and calling xml::xml_parse first. Use the longer form when the XML is a string inside other data, such as a field of a JSON document.
• For path-style queries such as //user[@role="dev"]/name, see Query XML with XPath. To go from XML to JSON and back, see Convert between XML and JSON.
• For RSS 2.0 and Atom feeds specifically, the feed.mq extension module parses both into one shape.