XPath Queries

Compile a Query

Use one of the xpath overloads on XMLDocument to compile an XPath expression and receive an XMLPath collection. XPath selects element nodes only, so there is no need to filter out text nodes:

if let titles = document.xpath("//title") {
    for title in titles {
        print(title.content)
    }
}

Use Predicates

Standard XPath predicate syntax narrows results. Attribute predicates are particularly useful:

// Select the book whose id attribute equals "2"
if let matches = document.xpath("//book[@id='2']") {
    for book in matches {
        print(book.attribute(named: "id") ?? "")
    }
}

Treat Results as a Collection

XMLPath conforms to RandomAccessCollection, so you can use familiar collection operations — subscripting, iteration, map, first, count — without converting the result set first:

if let books = document.xpath("//book") {
    print(books.count)           // number of results
    print(books[0].name)         // first result by index
    let titles = books.map { $0.children.first(where: { $0.name == "title" })?.content ?? "" }
}

Register Namespaces

When an XPath expression uses namespace prefixes, register them before evaluating. The overload that takes [(prefix: String, href: String)] tuples is the most direct approach when the namespace URIs are known at compile time:

let ns: [(prefix: String, href: String)] = [
    ("atom", "http://www.w3.org/2005/Atom")
]
if let titles = document.xpath("//atom:title", namespaces: ns) {
    for title in titles {
        print(title.content)
    }
}

The overload that accepts an AnySequence<XMLNameSpace> is convenient when the namespaces are read from the document itself:

let rootNS = document.rootElement.namespaces
if let titles = document.xpath("//atom:title", namespaces: rootNS) {
    for title in titles {
        print(title.content)
    }
}

See Working with Namespaces for a detailed treatment of namespace handling.