Any way to simplify the following? or reduce the boilerplate code with another function?
scala> val ns = <foo><bar id="1"><tag>one</tag><tag>uno</tag></bar><bar id="2"><tag>two</tag><tag>dos</tag></bar></foo> ns: scala.xml.Elem = <foo><bar id="1"><tag>one</tag><tag>uno</tag></bar><bar id="2"><tag>two</tag><tag>dos</tag></bar></foo> scala> (ns \\ "bar" filterNot{_ \\ "@id" find { _.text == "1" } isEmpty}) \\ "tag" res0: scala.xml.NodeSeq = NodeSeq(<tag>one</tag>, <tag>uno</tag>) I could only find a minor improvement, the find/isEmpty test can be replaced with exists:
(ns \\ "bar" filter { _ \\ "@id" exists (_.text == "1") }) \\ "tag" Edit after clarifying comment:
That's a really nice idea! Try this for size:
import xml._ implicit def richNodeSeq(ns: NodeSeq) = new { def \@(attribMatch: (String, String => Boolean)): NodeSeq = ns filter { _ \\ ("@" + attribMatch._1) exists (s => attribMatch._2(s.text)) } } ns \\ "bar" \@ ("id", _ == "1") \\ "tag" I used a predicate instead of hard-coding the attribute value comparison.
The best way is to define implicit class:
object XmlE { implicit class XmlE(val xml: NodeSeq) extends AnyVal { def \@(attr: (String, String => Boolean)): NodeSeq = { xml filter { _ \ ("@" + attr._1) exists (s => attr._2(s.text)) } } } } Then use it from another class :
import XmlE._ ..... 