ruby way之使用REXML解析xml

REXML 是一个完全用ruby写的processor ,他有多种api，其中两个经典的api是通过DOM-like 和SAX-like 来进行区分的。第一种是将整个文件读进内存，然后存储为一个分层的形式(也就是一棵树了).而第二种是"parse as you go",当你的文件很大，并且内存受到限制的时候，比较适合用这种。

看下面的book.xml:

引用

<library shelf="Recent Acquisitions">
    <section name="Ruby">
        <book isbn="0672328844">
        <title>The Ruby Way</title>
        <author>Hal Fulton</author>
        <description>
            Second edition. The book you are now reading.
            Ain't recursion grand?
        </description>
        </book>
    </section>
    <section name="Space">
        <book isbn="0684835509">
            <title>The Case for Mars</title>
            <author>Robert Zubrin</author>
            <description>Pushing toward a second home for the human
                race.
            </description>
        </book>
        <book isbn="074325631X">
            <title>First Man: The Life of Neil A. Armstrong</title>
            <author>James R. Hansen</author>
            <description>Definitive biography of the first man on
                the moon.
            </description>
        </book>
    </section>
</library>

1 Tree Parsing(也就是DOM-like)

我们需要require rexml/document 库，并且include REXML :

require 'rexml/document'
include REXML

input = File.new("books.xml")
doc = Document.new(input)

root = doc.root
puts root.attributes["shelf"]      # Recent Acquisitions

doc.elements.each("library/section") { |e| puts e.attributes["name"] }
# Output:
#   Ruby
#   Space

doc.elements.each("*/section/book") { |e| puts e.attributes["isbn"] }
# Output:
#   0672328844
#   0321445619
#   0684835509
#   074325631X

sec2 = root.elements[2]
author = sec2.elements[1].elements["author"].text       # Robert Zubrin

这里要注意的是xml中的属性和值被表示为一个hash，因此我们能够通过attributes[]来提取我们需要的值，元素的值还能通过类似于path的字符串或者整数来取得.其中用整数取的话，是1-based而不是0-based.

2  Stream Parsing（也就是SAX-like Parsing)

这边使用了一个小技巧，那就是定义了一个listener 类，它将会在parse的时候被回调:

require 'rexml/document'
require 'rexml/streamlistener'
include REXML

class MyListener
  include REXML::StreamListener
  def tag_start(*args)
    puts "tag_start: #{args.map {|x| x.inspect}.join(', ')}"
  end

  def text(data)
    return if data =~ /^\w*$/     # whitespace only
    abbrev = data[0..40] + (data.length > 40 ? "..." : "")
    puts "  text   :   #{abbrev.inspect}"
  end
end

list = MyListener.new
source = File.new "books.xml"
Document.parse_stream(source, list)

这里介绍一下StreamListener 模块，这个模块它提供了几个空的回调方法，因此你可以为了实现你自己的功能而覆盖它.当parser 进入一个tag时，就会调用tag_start方法.而text方法也是类似的，他只不过是当读取到数据时会被回调,它的输出是这样的：

tag_start: "library", {"shelf"=>"Recent Acquisitions"}
tag_start: "section", {"name"=>"Ruby"}
tag_start: "book", {"isbn"=>"0672328844"}
tag_start: "title", {}
  text   :   "The Ruby Way"
.........................................

3 XPath

REXML通过XPath 类来提供Xpath的支持. 它也同时支持DOM-like和SAX-like .还是前面的那个xml文件，我们使用Xpath可以这样做：

book1 = XPath.first(doc, "//book")    # Info for first book found
p book1

# Print out all titles
XPath.each(doc, "//title") { |e| puts e.text }

# Get an array of all of the "author" elements in the document.
names = XPath.match(doc, "//author").map {|x| x.text }
p names

输出是类似于下面的：

<book isbn='0672328844'> ... </>
The Ruby Way
The Case for Mars
First Man: The Life of Neil A. Armstrong
["Hal Fulton", "Robert Zubrin", "James R. Hansen"]