Clojure世界：XML处理-CSDN博客

XML处理也是个常见的编程工作，虽然说在Clojure里你很少使用XML做配置文件，但是跟遗留系统集成或者处理和其他系统通讯，可能都需要处理XML。

Clojure的标准库clojure.xml就是用来干这个事情的。一个简单的例子如下，首先我们要解析的是下面这个简单的XML：

<? xml version="1.0" encoding="UTF-8" ?>
< books >
 < book >
 < title > The joy of clojure </ title >
 < author > Michael Fogus / Chris House </ author >
 </ book >
 < book >
 < title > Programming clojure </ title >
 < author > Stuart Halloway </ author >
 </ book >
 < book >
 < title > Practical clojure </ title >
 < author > Luke Van der Hart </ author >
 </ book >
</ books >

解析xml用clojure.xml/parse方法即可，该方法返回一个clojure.xml/element这个struct-map组成的一棵树：

user => (use ' [clojure.xml])
nil
user => (parse " test.xml " )
{:tag :books, :attrs nil, :content
[{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " The joy of clojure " ]} {:tag :author, :attrs nil, :content [ " Michael Fogus / Chris House " ]}]}
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Programming clojure " ]} {:tag :author, :attrs nil, :content [ " Stuart Halloway " ]}]}
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Practical clojure " ]} {:tag :author, :attrs nil, :content [ " Luke Van der Hart " ]}]}]}

这是一个嵌套的数据结构，每个节点都是clojure.xml/element结构，element包括：

(defstruct element :tag :attrs :content)

tag、attrs和content属性，tag就是该节点的标签，attrs是一个属性的map，而content是它的内容或者子节点。element是一个struct map，它也定义了三个方法来分别获取这三个属性：

user => ( def x (parse " test.xml " ))
# 'user/x
user => (tag x)
:books
user => (attrs x)
nil
user => (content x)
[{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " The joy of clojure " ]} {:tag :author, :attrs nil, :content [ " Michael Fogus / Chris House " ]}]}
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Programming clojure " ]} {:tag :author, :attrs nil, :content [ " Stuart Halloway " ]}]}
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Practical clojure " ]} {:tag :author, :attrs nil, :content [ " Luke Van der Hart " ]}]}]

books节点是root node，它的content就是三个book子节点，子节点组织成一个vector，我们可以随意操作：

user => (tag (first (content x)))
:book
user => (content (first (content x)))
[{:tag :title, :attrs nil, :content [ " The joy of clojure " ]} {:tag :author, :attrs nil, :content [ " Michael Fogus / Chris House " ]}]
user => (content (first (content (first (content x)))))
[ " The joy of clojure " ]

额外提下，clojure.xml是利用SAX API做解析的。同样它还有个方法，可以将解析出来的结构还原成xml，通过emit：

user=> (emit x)

<? xml version='1.0' encoding='UTF-8' ?>
< books >
< book >
< title >
The joy of clojure
</ title >
< author >
Michael Fogus / Chris House
</ author >
</ book >
< book >

如果你要按照深度优先的顺序遍历xml，可以利用xml-seq将解析出来的树构成一个按照深度优先顺序排列节点的LazySeq，接下来就可以按照seq的方式处理，比如利用for来过滤节点：

user=> (for [node (xml-seq x)
:when (= :author (:tag node))]
(first (:content node)))
("Michael Fogus / Chris House" "Stuart Halloway" "Luke Van der Hart")

通过:when指定了条件，要求节点的tag是author，这样就可以查找出所有的author节点的content，是不是很方便？就像写英语描述一样。

更进一步，如果你想操作parse解析出来的这棵树，你还可以利用clojure.zip这个标准库，它有xml-zip函数将xml转换成zipper结构，并提供一系列方法来操作这棵树：

user=>(def xz (xml- zip x))
# 'user/xz
user=> (node (down xz))
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " The joy of clojure " ]} {:tag :author, :attrs nil, :content [ " Michael Fogus / Chris House " ]}]}
user => (-> xz down right node)
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Programming clojure " ]} {:tag :author, :attrs nil, :content [ " Stuart Halloway " ]}]}
user => (-> xz down right right node)
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Practical clojure " ]} {:tag :author, :attrs nil, :content [ " Luke Van der Hart " ]}]}
user => (-> xz down right right lefts)
({:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " The joy of clojure " ]} {:tag :author, :attrs nil, :content [ " Michael Fogus / Chris House " ]}]}
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Programming clojure " ]} {:tag :author, :attrs nil, :content [ " Stuart Halloway " ]}]})

是不是酷得一塌糊涂？可以通过up,down,left,right,lefts,rights,来查找节点的邻近节点，可以通过node来得到节点本身。一切显得那么自然。更进一步，你还可以“编辑“这棵树，比如删除The joy of clojure这本书：

user => ( def loc - in - new - tree (remove (down xz)))
# 'user/loc-in-new-tree
user => (root loc - in - new - tree)
{:tag :books, :attrs nil, :content
[{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Programming clojure " ]} {:tag :author, :attrs nil, :content [ " Stuart Halloway " ]}]}
{:tag :book, :attrs nil, :content [{:tag :title, :attrs nil, :content [ " Practical clojure " ]} {:tag :author, :attrs nil, :content [ " Luke Van der Hart " ]}]}]}

ok,只剩下两本书了，更多方法还包括replace做替换，edit更改节点等。因此编辑XML并重新生成，你一般可以利用clojure.zip来更改树，最后利用clojure.xml/emit将更改还原为xml。

生成xml除了emit方法，还有一个contrib库，也就是 prxml，这个库的clojure 1.3版本有人维护了一个分支，在这里。主要方法就是prxml，它可以将clojure的数据结构转换成xml：