# 9-I/O
[原文链接](http://code.google.com/p/guava-libraries/wiki/IOExplained) [译文链接](http://ifeve.com/google-guava-io) **译者：**沈义扬
## 字节流和字符流
Guava使用术语”流” 来表示可关闭的，并且在底层资源中有位置状态的I/O数据流。术语”字节流”指的是InputStream或OutputStream，”字符流”指的是Reader 或Writer（虽然他们的接口Readable 和Appendable被更多地用于方法参数）。相应的工具方法分别在[`ByteStreams`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/ByteStreams.html) 和[`CharStreams`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/CharStreams.html)中。
大多数Guava流工具一次处理一个完整的流，并且/或者为了效率自己处理缓冲。还要注意到，接受流为参数的Guava方法不会关闭这个流：关闭流的职责通常属于打开流的代码块。
其中的一些工具方法列举如下：
| **ByteStreams** | **CharStreams** |
|:--- |:--- |
| [`byte[] toByteArray(InputStream)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/ByteStreams.html#toByteArray%28java.io.InputStream%29) | [`String toString(Readable)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/CharStreams.html#toString%28java.lang.Readable%29) |
| N/A | [`List<String> readLines(Readable)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharStreams.html#readLines%28java.lang.Readable%29) |
| [`long copy(InputStream, OutputStream)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/ByteStreams.html#copy%28java.io.InputStream, java.io.OutputStream%29) | [`long copy(Readable, Appendable)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/CharStreams.html#copy%28java.lang.Readable, java.lang.Appendable%29) |
| [`void readFully(InputStream, byte[])`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/ByteStreams.html#readFully%28java.io.InputStream, byte[]%29) | N/A |
| [`void skipFully(InputStream, long)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteStreams.html#skipFully%28java.io.InputStream, long%29) | [`void skipFully(Reader, long)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharStreams.html#skipFully%28java.io.Reader, long%29) |
| [`OutputStream nullOutputStream()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteStreams.html#nullOutputStream%28%29) | [`Writer nullWriter()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharStreams.html#nullWriter%28%29) |
**关于InputSupplier 和OutputSupplier要注意：**
在ByteStreams、CharStreams以及com.google.common.io包中的一些其他类中，某些方法仍然在使用InputSupplier和OutputSupplier接口。这两个借口和相关的方法是不推荐使用的：它们已经被下面描述的source和sink类型取代了，并且最终会被移除。
## 源与汇
通常我们都会创建I/O工具方法，这样可以避免在做基础运算时总是直接和流打交道。例如，Guava有Files.toByteArray(File) 和Files.write(File, byte[])。然而，流工具方法的创建经常最终导致散落各处的相似方法，每个方法读取不同类型的源
或写入不同类型的汇[sink]。例如，Guava中的Resources.toByteArray(URL)和Files.toByteArray(File)做了同样的事情，只不过数据源一个是URL，一个是文件。
为了解决这个问题，Guava有一系列关于源与汇的抽象。源或汇指某个你知道如何从中打开流的资源，比如File或URL。源是可读的，汇是可写的。此外，源与汇按照字节和字符划分类型。
| **字节** | **字符** |
|:--- |:--- |
| **读** | [`ByteSource`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/ByteSource.html) | [`CharSource`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/CharSource.html) |
| **写** | [`ByteSink`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/ByteSink.html) | [`CharSink`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/CharSink.html) |
源与汇API的好处是它们提供了通用的一组操作。比如，一旦你把数据源包装成了ByteSource，无论它原先的类型是什么，你都得到了一组按字节操作的方法。
### 创建源与汇
Guava提供了若干源与汇的实现：
| **字节** | **字符** |
|:--- |:--- |
| [`Files.asByteSource(File)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Files.html#asByteSource%28java.io.File%29) | [`Files.asCharSource(File, Charset)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Files.html#asCharSource%28java.io.File, java.nio.charset.Charset%29) |
| [`Files.asByteSink(File, FileWriteMode...)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Files.html#asByteSink%28java.io.File, com.google.common.io.FileWriteMode...%29) | [`Files.asCharSink(File, Charset, FileWriteMode...)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Files.html#asCharSink%28java.io.File, java.nio.charset.Charset, com.google.common.io.FileWriteMode...%29) |
| [`Resources.asByteSource(URL)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Resources.html#asByteSource%28java.net.URL%29) | [`Resources.asCharSource(URL, Charset)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Resources.html#asCharSource%28java.net.URL, java.nio.charset.Charset%29) |
| [`ByteSource.wrap(byte[])`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#wrap%28byte[]%29) | [`CharSource.wrap(CharSequence)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#wrap%28java.lang.CharSequence%29) |
| [`ByteSource.concat(ByteSource...)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#concat%28com.google.common.io.ByteSource...%29) | [`CharSource.concat(CharSource...)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#concat%28com.google.common.io.CharSource...%29) |
| [`ByteSource.slice(long, long)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#slice%28long, long%29) | N/A |
| N/A | [`ByteSource.asCharSource(Charset)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#asCharSource%28java.nio.charset.Charset%29) |
| N/A | [`ByteSink.asCharSink(Charset)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSink.html#asCharSink%28java.nio.charset.Charset%29) |
此外，你也可以继承这些类，以创建新的实现。
注：把已经打开的流（比如InputStream）包装为源或汇听起来是很有诱惑力的，但是应该避免这样做。源与汇的实现应该在每次openStream()方法被调用时都创建一个新的流。始终创建新的流可以让源或汇管理流的整个生命周期，并且让多次调用openStream()返回的流都是可用的。此外，如果你在创建源或汇之前创建了流，你不得不在异常的时候自己保证关闭流，这压根就违背了发挥源与汇API优点的初衷。
### 使用源与汇
一旦有了源与汇的实例，就可以进行若干读写操作。
**通用操作**
所有源与汇都有一些方法用于打开新的流用于读或写。默认情况下，其他源与汇操作都是先用这些方法打开流，然后做一些读或写，最后保证流被正确地关闭了。这些方法列举如下：
* openStream()：根据源与汇的类型，返回InputStream、OutputStream、Reader或者Writer。
* openBufferedStream()：根据源与汇的类型，返回InputStream、OutputStream、BufferedReader或者BufferedWriter。返回的流保证在必要情况下做了缓冲。例如，从字节数组读数据的源就没有必要再在内存中作缓冲，这就是为什么该方法针对字节源不返回BufferedInputStream。字符源属于例外情况，它一定返回BufferedReader，因为BufferedReader中才有readLine()方法。
**源操作**
| **字节源** | **字符源** |
|:--- |:--- |
| [`byte[] read()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#read%28%29) | [`String read()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#read%28%29) |
| N/A | [`ImmutableList<String> readLines()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#readLines%28%29) |
| N/A | [`String readFirstLine()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#readFirstLine%28%29) |
| [`long copyTo(ByteSink)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#copyTo%28com.google.common.io.ByteSink%29) | [`long copyTo(CharSink)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#copyTo%28com.google.common.io.CharSink%29) |
| [`long copyTo(OutputStream)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#copyTo%28java.io.OutputStream%29) | [`long copyTo(Appendable)` ](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#copyTo%28java.lang.Appendable%29) |
| [`long size()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#size%28%29) (in bytes) | N/A |
| [`boolean isEmpty()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#isEmpty%28%29) | [`boolean isEmpty()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSource.html#isEmpty%28%29) |
| [`boolean contentEquals(ByteSource)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#contentEquals%28com.google.common.io.ByteSource%29) | N/A |
| [`HashCode hash(HashFunction)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSource.html#hash%28com.google.common.hash.HashFunction%29) | N/A |
汇操作
| **字节汇** | **字符汇** |
|:--- |:--- |
| [`void write(byte[])`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSink.html#write%28byte[]%29) | [`void write(CharSequence)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSink.html#write%28java.lang.CharSequence%29) |
| [`long writeFrom(InputStream)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/ByteSink.html#writeFrom%28java.io.InputStream%29) | [`long writeFrom(Readable)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSink.html#writeFrom%28java.lang.Readable%29) |
| N/A | [`void writeLines(Iterable<? extends CharSequence>)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSink.html#writeLines%28java.lang.Iterable%29) |
| N/A | [`void writeLines(Iterable<? extends CharSequence>, String)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/CharSink.html#writeLines%28java.lang.Iterable, java.lang.String%29) |
### 范例
```
//Read the lines of a UTF-8 text file
ImmutableList<String> lines = Files.asCharSource(file, Charsets.UTF_8).readLines();
//Count distinct word occurrences in a file
Multiset<String> wordOccurrences = HashMultiset.create(
Splitter.on(CharMatcher.WHITESPACE)
.trimResults()
.omitEmptyStrings()
.split(Files.asCharSource(file, Charsets.UTF_8).read()));
//SHA-1 a file
HashCode hash = Files.asByteSource(file).hash(Hashing.sha1());
//Copy the data from a URL to a file
Resources.asByteSource(url).copyTo(Files.asByteSink(file));
```
## 文件操作
除了创建文件源和文件的方法，Files类还包含了若干你可能感兴趣的便利方法。
| [`createParentDirs(File)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/Files.html#createParentDirs%28java.io.File%29) | 必要时为文件创建父目录 |
|:--- |:--- |
| [`getFileExtension(String)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/Files.html#getFileExtension%28java.lang.String%29) | 返回给定路径所表示文件的扩展名 |
| [`getNameWithoutExtension(String)`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Files.html#getNameWithoutExtension%28java.lang.String%29) | 返回去除了扩展名的文件名 |
| [`simplifyPath(String)`](http://docs.guava-libraries.googlecode.com/git-history/release/javadoc/com/google/common/io/Files.html#simplifyPath%28java.lang.String%29) | 规范文件路径，并不总是与文件系统一致，请仔细测试 |
| [`fileTreeTraverser()`](http://docs.guava-libraries.googlecode.com/git/javadoc/com/google/common/io/Files.html#fileTreeTraverser%28%29) | 返回TreeTraverser用于遍历文件树 |
- Google Guava官方教程（中文版）
- 1-基本工具
- 1.1-使用和避免null
- 1.2-前置条件
- 1.3-常见Object方法
- 1.4-排序: Guava强大的”流畅风格比较器”
- 1.5-Throwables：简化异常和错误的传播与检查
- 2-集合
- 2.1-不可变集合
- 2.2-新集合类型
- 2.3-强大的集合工具类：java.util.Collections中未包含的集合工具
- 2.4-集合扩展工具类
- 3-缓存
- 4-函数式编程
- 5-并发
- 5.1-google Guava包的ListenableFuture解析
- 5.2-Google-Guava Concurrent包里的Service框架浅析
- 6-字符串处理：分割，连接，填充
- 7-原生类型
- 9-I/O
- 10-散列
- 11-事件总线
- 12-数学运算
- 13-反射