使用的 SAX解析器。 SAX解析器设计用于处理巨大的XML文件。相反，在加载XML文件到内存中一气呵成的，它走在文档元素乘元素，并通知您。

Use the SAX parser. SAX parsers are designed to handle huge XML files. Instead of loading the XML file into memory in one go, it walks over the document element-by-element and notifies you.

此外，如果XML文件是非常大的，你可能还需要看看文件的加载方式。不要打开该文件，并全部内容送入SAX解析器一气呵成。相反，阅读块逐块（每次例如4KB块），并养活到SAX解析器。

Additionally, if the XML file is really big, you may also want to look at how the file is loaded. Don't open the file and feed the entire contents into the SAX parser in one go. Instead, read it chunk-by-chunk (e.g. 4Kb blocks at a time) and feed that into the SAX parser.

编辑：SAX解析器的作品非常不同，从一个DOM解析器。基本上，它只是通过一次原稿一种元素。每当找到一个打开或关闭的标签，它调用你的函数（如回调），并告诉它的标签是什么，是什么数据（如果有的话）。它在一开始并贯穿始终的推移，永不返回。这是连载。这意味着两件事情：

A SAX parser works very differently from a DOM parser. Basically, it just goes through the document one element at a time. Whenever it finds an open or close tag, it calls one of your functions (as a callback) and tells it what the tag is and what the data is (if any). It starts at the beginning and goes through to the end and never goes back. It's serial. This means two things:

• 更​​多code。您的回调需要确定在遇到某些标记时做什么，应该忽略什么标签，等等。 SAX解析器不回去，所以如果你需要记住的东西后，你需要做的一切你自己。所以，是的，这将是更多的工作来处理包含许多不同的标记许多API。

• More code. Your callback needs to determine what to do when certain tags are encountered, what tags should be skipped, etcetera. A SAX parser doesn't go back, so if you need to remember anything for later, you need to do that all yourself. So yeah, it will be more work to deal with many APIs containing many different tags.

这可以解析XML部分。它并不关心你养活，如果只是第4 KB的XML文件。它不会产生一个错误，而只是要求另一块数据时，它的完成。只有当它遇到一个不匹配的结束标记（或者你停得太早喂养它的数据），它会产生一个错误。

It can parse partial XML. It doesn't care that you feed if just the first 4 Kb of an XML file. It will not generate an error but simply ask for another chunk data when it's done. Only when it encounters a mismatched closing tag (or you stop feeding it data too soon) will it generate an error.

所以，是的，这是更多的工作。但回报是更大的速度，没有问题的解析巨大的文件将不适合到内存中。

So yeah, it's more work. But the payoff is much greater speed and no problem parsing huge files that would not fit into memory.