java io 流
io 流总览
io.png
io 流主要提供四个接口
-
InputStream: 输入字节流 -
OutputStream: 输出字节流 -
Reader: 输入字符流 -
Writer: 输出字符流
InputStream
InputStream 输入字节流，关注字节的读取，io 包提供如下 InputStream 的实现
-
ByteArrayInputStream: 字节数组输入流 -
FileInputStream: 文件字节输入流 -
PipedInputStream: 管道输入流，可和其他的 PipedOutStream 连接，通常用于线程间通信 -
DataInputStream: 二进制数据输入流 -
ObjectInputStream: 对象输入流 -
BufferedInputStream: 带缓冲 buffer 的字节输入流 -
SequenceInputStream: 能将多个字节流合并成一个 -
PushbackInputStream: 能回退的字节流
InputStream 提供如下接口:
-
read: 从流中读取一个字节 -
read(buffer): 从流中读取字节到 buffer 中，返回真实读取的字节数 -
read(buffer, offset, length): 从流中读取 length 个字节，写入到 buffer 的 offset 处，返回真实读取的字节数 -
readNBytes(buffer, offset, length): 和 read 一样，但是保证读取 length 个字节，除非流中没有数据 -
readAllBytes: 读取所有字节，返回一个字节数组 -
skip: 跳过前 n 个字节 -
available: 剩余字节数 -
mark: 标记当前读取的位置 -
reset: 将流指针重置到上次标记的位置 -
close: 关闭流，释放资源
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
assertEquals(in.read(), '0');
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
byte[] buf = new byte[4];
assertEquals(in.read(buf), 4);
assertArrayEquals(buf, "0123".getBytes());
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
byte[] buf = new byte[20];
assertEquals(in.read(buf), 10);
assertArrayEquals(Arrays.copyOf(buf, 10), "0123456789".getBytes());
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
byte[] buf = new byte[20];
assertEquals(in.read(buf, 1, 4), 4);
assertArrayEquals(Arrays.copyOfRange(buf, 1, 1 + 4), "0123".getBytes());
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
byte[] buf = new byte[20];
assertEquals(in.readNBytes(buf, 1, 4), 4);
assertArrayEquals(Arrays.copyOfRange(buf, 1, 1 + 4), "0123".getBytes());
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
assertArrayEquals(in.readAllBytes(), "0123456789".getBytes());
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
assertEquals(in.skip(2), 2);
assertEquals(in.available(), 8);
assertEquals(in.read(), '2');
assertEquals(in.available(), 7);
in.mark(0);
assertEquals(in.read(), '3');
in.reset();
assertEquals(in.available(), 7);
assertEquals(in.read(), '3');
in.close();
}
{
InputStream in = new ByteArrayInputStream("0123456789".getBytes());
for (int ch = in.read(); ch != -1; ch = in.read()) {
System.out.println(ch);
}
}
OutputStream
OutputStream 输出字节流，关注字节的写入，io 包提供了如下 OutputStream 的实现
-
ByteArrayOutputStream: 输出 byte 数组 -
FileOutputStream: 文件输出流 -
PipedOutputStream: 管道输出流，可和其他的 PipedInputStream 连接，通常用于线程间通信 -
DataOutputStream: 二进制数据输出流 -
ObjectOutputStream: 对象输出流 -
BufferedOutputStream: 带缓冲 buffer 的输出流 -
SequenceOutputStream: 能将多个输出流合并成一个
OutputStream 提供如下接口:
-
write: 写入一个字节 -
write(buffer): 写入 buffer 中的数据 -
write(buffer, offset, length): 写入 buffer 从 offset 起的 length 个字节的数据 -
flush: 将缓冲区的数据刷到实际的存储中 -
close: 关闭流
OutputStream out = new ByteArrayOutputStream();
out.write('0');
out.write("123456789".getBytes());
out.write("0123456789".getBytes(), 1, 2);
out.flush();
out.close();
Reader
Reader 字符输入流，关注字符的读取，io 包提供如下 Reader 的实现
-
CharArrayReader: 字符数组输入流 -
FileReader: 文件字符输入流 -
PipedReader: 管道输入流，可以和PipedWriter连接，通常用于线程间通信 -
StringReader: 字符串输入流 -
BufferedReader: 带缓冲 buffer 的字符输入流 -
LineNumberReader: 带行号的字符输入流 -
PushbackReader: 能回退的字符输入流 -
InputStreamReader: 输入字节流转字符流
Reader 提供如下接口:
-
read: 从流中读取一个字符 -
read(buffer): 从流中读取字符到 buffer 中，返回真实读取的字符数 -
read(buffer, offset, length): 从流中读取 length 个字符，写入到 buffer 的 offset 处，返回真实读取的字符数 -
read(CharBuffer: 从流中读取字符到 CharBuffer 中，返回真实读取的字符数 -
skip: 跳过前 n 个字符 -
mark: 标记当前读取的位置 -
reset: 将流指针重置到上次标记的位置 -
close: 关闭流，释放资源
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
assertEquals(reader.read(), '0');
}
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
char[] buf = new char[4];
assertEquals(reader.read(buf), 4);
assertArrayEquals(buf, "0123".toCharArray());
}
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
char[] buf = new char[20];
assertEquals(reader.read(buf), 10);
assertArrayEquals(Arrays.copyOf(buf, 10), "0123456789".toCharArray());
}
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
char[] buf = new char[20];
assertEquals(reader.read(buf, 1, 4), 4);
assertArrayEquals(Arrays.copyOfRange(buf, 1, 1 + 4), "0123".toCharArray());
}
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
CharBuffer buf = CharBuffer.allocate(20);
assertEquals(reader.read(buf), 10);
}
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
assertTrue(reader.ready());
assertEquals(reader.skip(2), 2);
assertEquals(reader.read(), '2');
reader.mark(0);
assertEquals(reader.read(), '3');
reader.reset();
assertEquals(reader.read(), '3');
reader.close();
}
{
Reader reader = new CharArrayReader("0123456789".toCharArray());
for (int ch = reader.read(); ch != -1; ch = reader.read()) {
System.out.println(ch);
}
}
Writer
Writer 字符输出流，关注字符的写入，io 包提供如下 Writer 的实现
-
CharArrayWriter: 字符数组输出流 -
FileWriter: 文件字符输出流 -
PipedWriter: 管道输出流，可以和PipedReader连接，通常用于线程间通信 -
StringWriter: 字符串输出流 -
BufferedWriter: 带缓冲 buffer 的字符输出流 -
OutputStreamWriter: 输出字节流转字符流
Writer 提供如下接口:
-
write(char): 写入一个字符 -
write(string): 写入一个字符串 -
write(string, offset, length): 写入 string 从 offset 起的 length 个字符的数据 -
write(char[]): 写入字符数组中的数据 -
write(char[], offset, length): 写入字符数组从 offset 起的 length 个字符的数据 -
append(ch): 写入一个字符，和 write 一样 -
append(CharSequence): 写入字符序列的所有数据(String, StringBuilder, StringBuffer 都是 CharSequence 的子类) -
append(CharSequence, offset, length): 写入字符序列从 offset 起的 length 个字符的数据 -
flush: 将缓冲区的数据刷到实际的存储中 -
close: 关闭流
Writer writer = new CharArrayWriter();
writer.write('0');
writer.write("0123456789");
writer.write("0123456789", 1, 4);
writer.write("0123456789".toCharArray());
writer.write("0123456789".toCharArray(), 1, 4);
writer.append('0');
writer.append(new StringBuilder("0123456789"));
writer.append(new StringBuilder("0123456789"), 1, 4);
writer.flush();
writer.close();
文件字节流
文件字节流关注文件的读取和写入
{
FileOutputStream fout = new FileOutputStream("/tmp/test.txt");
fout.write("No patient who, who has no wisdom".getBytes());
fout.close();
}
{
FileInputStream fin = new FileInputStream("/tmp/test.txt");
assertArrayEquals(fin.readAllBytes(), "No patient who, who has no wisdom".getBytes());
fin.close();
}
缓冲字节流
缓冲字节流采用装饰者模式，装饰在其他流上，使流拥有了缓存功能，从而提高读写了效率
{
BufferedOutputStream bout = new BufferedOutputStream(new FileOutputStream("/tmp/test.txt"));
bout.write("People lack the willpower, rather than strength".getBytes());
bout.close();
}
{
BufferedInputStream bin = new BufferedInputStream(new FileInputStream("/tmp/test.txt"));
assertArrayEquals(bin.readAllBytes(), "People lack the willpower, rather than strength".getBytes());
bin.close();
}
二进制字节流
二进制字节流关注在基本数据类型的读取和写入，采用装饰者模式，能装饰在其他流上
DataOutputStream 在 OutputStream 的基础上新增了如下接口:
-
writeBoolean: 写入一个 boolean 值 -
writeByte: 写入一个字节 -
writeShort: 写入一个短整型 -
writeInt: 写入一个整型 -
writeLong: 写入一个长整型 -
writeFloat: 写入一个浮点型 -
writeDouble: 写入一个双精度浮点型 -
writeChar: 写入一个字符 -
writeUTF: 写入一个 unicode 字符串
DataInputStream 在 InputStream 的基础上新增了如下接口:
-
readBoolean: 读取一个 boolean 值 -
readByte: 读取一个字节 -
readShort: 读取一个 short -
readInt: 读取一个整型 -
readLong: 读取一个长整型 -
readFloat: 读取一个浮点型 -
readDouble: 读取一个双精度浮点型 -
readChar: 读取一个字符 -
readUTF: 读取一个 unicode 字符串
{
DataOutputStream dout = new DataOutputStream(new BufferedOutputStream(new FileOutputStream("/tmp/test.txt")));
dout.writeBoolean(false);
dout.writeByte('x');
dout.writeShort(123);
dout.writeInt(123456);
dout.writeLong(123456789);
dout.writeFloat((float) 123.456);
dout.writeDouble(123.456);
dout.writeUTF("Rome wasn’t built in one day");
dout.close();
}
{
DataInputStream din = new DataInputStream(new BufferedInputStream(new FileInputStream("/tmp/test.txt")));
assertEquals(din.readBoolean(), false);
assertEquals(din.readByte(), 'x');
assertEquals(din.readShort(), 123);
assertEquals(din.readInt(), 123456);
assertEquals(din.readLong(), 123456789);
assertEquals(din.readFloat(), (float) 123.456);
assertEquals(din.readDouble(), 123.456);
assertEquals(din.readUTF(), "Rome wasn’t built in one day");
din.close();
}
对象字节流
对象字节流关注对象的写入和读取，同时拥有二进制字节流的所有功能，同样采用装饰者模式
ObjectOutputStream 相比 DataOutputStream 新增了如下接口:
-
writeObject: 写入任何 Serializable 对象
ObjectInputStream 相比 DataInputStream 新增了如下接口:
-
readObject: 从流中读取一个对象
{
ObjectOutputStream oout = new ObjectOutputStream(new BufferedOutputStream(new FileOutputStream("/tmp/test.txt")));
oout.writeBoolean(false);
oout.writeByte('x');
oout.writeShort(123);
oout.writeInt(123456);
oout.writeLong(123456789);
oout.writeFloat((float) 123.456);
oout.writeDouble(123.456);
oout.writeUTF("Nothing is impossible to a willing heart");
oout.writeObject(new Point(123, 456));
oout.close();
}
{
ObjectInputStream oin = new ObjectInputStream(new BufferedInputStream(new FileInputStream("/tmp/test.txt")));
assertEquals(oin.readBoolean(), false);
assertEquals(oin.readByte(), 'x');
assertEquals(oin.readShort(), 123);
assertEquals(oin.readInt(), 123456);
assertEquals(oin.readLong(), 123456789);
assertEquals(oin.readFloat(), (float) 123.456);
assertEquals(oin.readDouble(), 123.456);
assertEquals(oin.readUTF(), "Nothing is impossible to a willing heart");
Point point = (Point) oin.readObject();
assertEquals(point.x, 123);
assertEquals(point.y, 456);
oin.close();
}
回退字节流
可回退字节流内部维护了一个固定大小的缓冲区(可通过构造函数配置 buffer 的大小)，允许将字节回退到缓冲区，如果超过了缓冲区大小，会抛出异常
PushbackInputStream 在 InputStream 的基础上新增了如下接口:
-
unread: 回退一个字节 -
unread(buffer): 将 buffer 中的数据回退到流的缓冲区 -
unread(buffer, offset, length): 从 buffer 的 offset 处回退 length 个字节到流缓冲区
PushbackInputStream pin = new PushbackInputStream(new ByteArrayInputStream("Failure is the mother of success".getBytes()), 10);
byte[] buf = new byte[7];
assertEquals(pin.read(buf), 7);
assertArrayEquals(buf, "Failure".getBytes());
pin.unread(buf);
assertEquals(pin.read(buf), 7);
assertArrayEquals(buf, "Failure".getBytes());
// 超过 buffer 的大小，抛出 IOException
assertThrows(IOException.class, () -> pin.unread("01234567890".getBytes()));
SequenceInputStream
SequenceInputStream 将多个 InputStream 合并成一个
InputStream in1 = new ByteArrayInputStream("For man is man and master of his fate\n".getBytes());
InputStream in2 = new ByteArrayInputStream("Cease to struggle and you cease to live\n".getBytes());
Vector<InputStream> vi = new Vector<>(List.of(in1, in2));
SequenceInputStream sin = new SequenceInputStream(vi.elements());
assertArrayEquals(sin.readAllBytes(), "For man is man and master of his fate\nCease to struggle and you cease to live\n".getBytes());
管道字节流
PipedInputStream 和 PipedOutputStream 通过调用 connect 方法建立连接，往 PipedOutputStream 写入，能从 PipedInputStream 读取，这种管道模式是一对一的，对一个管道流建立两次连接会抛出异常
PipedOutputStream 在 OutputStream 的基础上提供如下接口:
-
connect: 与一个PipedInputStream建立连接，如果已经建立连接，将抛出异常
PipedInputStream 在 InputStream 的基础上提供如下接口:
-
connect: 与一个PipedOutputStream建立连接，如果已经建立连接，将抛出异常
ExecutorService es = Executors.newCachedThreadPool();
PipedInputStream pin = new PipedInputStream();
PipedOutputStream pout = new PipedOutputStream();
pin.connect(pout);
es.execute(() -> {
try {
ObjectOutputStream oout = new ObjectOutputStream(pout);
oout.writeInt(123456);
oout.writeUTF("如果你还没能找到让自己热爱的事业，继续寻找，不要放弃");
oout.close();
} catch (IOException e) {
e.printStackTrace();
}
});
es.execute(() -> {
try {
ObjectInputStream oin = new ObjectInputStream(pin);
assertEquals(oin.readInt(), 123456);
assertEquals(oin.readUTF(), "如果你还没能找到让自己热爱的事业，继续寻找，不要放弃");
oin.close();
} catch (IOException e) {
e.printStackTrace();
}
});
try {
es.shutdown();
while (!es.awaitTermination(1000, TimeUnit.MILLISECONDS)) {
// nothing to do
}
} catch (Exception e) {
e.printStackTrace();
}
文件字符流
文件字符流关注文件的读取和写入，使用默认的 utf-8 来编码
{
FileWriter fw = new FileWriter("/tmp/test.txt");
assertEquals(fw.getEncoding(), "UTF8");
System.out.println(fw.getEncoding());
fw.write("初学者的心态；拥有初学者的心态是件了不起的事情");
fw.flush();
fw.close();
}
{
FileReader fr = new FileReader("/tmp/test.txt");
assertEquals(fr.getEncoding(), "UTF8");
StringBuilder sb = new StringBuilder();
for (int ch = fr.read(); ch != -1; ch = fr.read()) {
sb.append((char) ch);
}
assertEquals(sb.toString(), "初学者的心态；拥有初学者的心态是件了不起的事情");
fr.close();
}
缓冲字符流
采用装饰者模式，装饰在其他字符流上，增加缓存功能，提高读写性能。Files 提供了缓冲字符流的构造，可以指定编码
BufferedWriter 在 Writer 的基础上，新增了如下接口:
-
newLine: 写入一个换行符
BufferedReader 在 Reader 的基础上，新增了如下接口:
-
readLine: 读取一个行，如果没有新的行，返回 null -
lines: 返回一个java.util.stream.Stream，支持 java 8 的流式处理
{
// BufferedWriter bw = new BufferedWriter(new FileWriter("/tmp/test.txt"));
BufferedWriter bw = Files.newBufferedWriter(Paths.get("/tmp/test.txt"), Charsets.UTF_8);
bw.write("穷则独善其身，达则兼济天下");
bw.newLine();
bw.write("玉不琢、不成器，人不学、不知义");
bw.newLine();
bw.close();
}
{
// BufferedReader br = new BufferedReader(new FileReader("/tmp/test.txt"));
BufferedReader br = Files.newBufferedReader(Paths.get("/tmp/test.txt"), Charsets.UTF_8);
assertEquals(br.readLine(), "穷则独善其身，达则兼济天下");
assertEquals(br.readLine(), "玉不琢、不成器，人不学、不知义");
assertEquals(br.readLine(), null);
br.close();
}
{
// BufferedReader br = new BufferedReader(new FileReader("/tmp/test.txt"));
BufferedReader br = Files.newBufferedReader(Paths.get("/tmp/test.txt"), Charsets.UTF_8);
assertThat(br.lines().collect(Collectors.toList()), equalTo(List.of(
"穷则独善其身，达则兼济天下",
"玉不琢、不成器，人不学、不知义"
)));
br.close();
}
StreamReaderWriter
InputStreamReader 和 OutputStreamWriter 能将字节流转化字符流，还可以指定编码
{
OutputStreamWriter ow = new OutputStreamWriter(new FileOutputStream("/tmp/test.txt"), "utf-8");
ow.write("你究竟是想一辈子卖糖水，还是希望获得改变世界的机遇");
ow.flush();
ow.close();
}
{
InputStreamReader rw = new InputStreamReader(new FileInputStream("/tmp/test.txt"), "utf-8");
StringBuilder sb = new StringBuilder();
for (int ch = rw.read(); ch != -1; ch = rw.read()) {
sb.append((char) ch);
}
assertEquals(sb.toString(), "你究竟是想一辈子卖糖水，还是希望获得改变世界的机遇");
rw.close();
}
字符串流
字符串构建的流
{
StringWriter sw = new StringWriter();
sw.write("学而不思则罔，思而不学则殆");
assertEquals(sw.getBuffer().toString(), "学而不思则罔，思而不学则殆");
sw.close();
}
{
StringReader sr = new StringReader("一年之计在于春，一日之计在于晨");
StringBuilder sb = new StringBuilder();
for (int ch = sr.read(); ch != -1; ch = sr.read()) {
sb.append((char) ch);
}
assertEquals(sb.toString(), "一年之计在于春，一日之计在于晨");
}
LineNumberReader
LineNumberReader 支持行号的字符流
LineNumberReader 在 Reader 的基础上，新增了如下接口:
-
setLineNumber: 设置开始的文件行号，默认是 1 -
getLineNumber: 获取当前的文件行号
{
BufferedWriter bw = new BufferedWriter(new FileWriter("/tmp/test.txt"));
bw.write("富贵不能淫\n贫贱不能移\n威武不能屈\n此之谓大丈夫\n");
bw.close();
}
{
LineNumberReader lr = new LineNumberReader(new BufferedReader(new FileReader("/tmp/test.txt")));
List<String> lines = new LinkedList<>();
for (String line = lr.readLine(); line != null; line = lr.readLine()) {
lines.add(lr.getLineNumber() + " " + line);
}
assertThat(lines, equalTo(List.of(
"1 富贵不能淫", "2 贫贱不能移", "3 威武不能屈", "4 此之谓大丈夫"
)));
}
回退字符流
可回退字符流内部维护了一个固定大小的缓冲区(可通过构造函数配置 buffer 的大小)，允许将字符回退到缓冲区，如果超过了缓冲区大小，会抛出异常
PushbackReader 在 Reader 的基础上新增了如下接口:
-
unread: 回退一个字符 -
unread(cbar[]): 将 buffer 中的数据回退到流的缓冲区 -
unread(char[], offset, length): 从 buffer 的 offset 处回退 length 个字节到流缓冲区
PushbackReader pr = new PushbackReader(new StringReader("蚍蜉撼大树,可笑不自量"), 10);
char[] buf = new char[5];
assertEquals(pr.read(buf), 5);
assertArrayEquals(buf, "蚍蜉撼大树".toCharArray());
pr.unread(buf);
assertEquals(pr.read(buf), 5);
assertArrayEquals(buf, "蚍蜉撼大树".toCharArray());
// 超过 buffer 的大小，抛出 IOException
assertThrows(IOException.class, () -> pr.unread("01234567890".toCharArray()));
管道字符流
PipedReader 和 PipedWriter 通过调用 connect 方法建立连接，往 PipedWriter 写入，能从 PipedReader 读取，这种管道模式是一对一的，对一个管道流建立两次连接会抛出异常
PipedWriter 在 Writer 的基础上提供如下接口:
-
connect: 与一个PipedReader建立连接，如果已经建立连接，将抛出异常
PipedReader 在 Reader 的基础上提供如下接口:
-
connect: 与一个PipedWriter建立连接，如果已经建立连接，将抛出异常
ExecutorService es = Executors.newCachedThreadPool();
PipedReader pr = new PipedReader();
PipedWriter pw = new PipedWriter();
pr.connect(pw);
es.execute(() -> {
try {
BufferedWriter bw = new BufferedWriter(pw);
bw.write("活着就是为了改变世界，难道还有其他原因吗");
bw.close();
} catch (Exception e) {
e.printStackTrace();
}
});
es.execute(() -> {
try {
BufferedReader br = new BufferedReader(pr);
assertEquals(br.readLine(), "活着就是为了改变世界，难道还有其他原因吗");
br.close();
} catch (Exception e) {
e.printStackTrace();
}
});
try {
es.shutdown();
while (!es.awaitTermination(1000, TimeUnit.MILLISECONDS)) {
// nothing to do
}
} catch (Exception e) {
e.printStackTrace();
}
