• QQ
  • nahooten@sina.com
  • 常州市九洲新世界花苑15-2

技术天地

Java中将InputStream读取为String

原创内容,转载请注明原文网址:http://homeqin.cn/a/wenzhangboke/jishutiandi/2019/0519/501.html

 

以下一共存在11种完成方式及其对应的性能测试后果:

1. 
应用IOUtils.toString (Apache Utils)

String result = IOUtils.toString(inputStream, StandardCharsets.UTF_8);

2. 应用CharStreams (guava)

String result = CharStreams.toString(new InputStreamReader(inputStream, Charsets.UTF_8));

3. 应用Scanner (JDK)

Scanner s = new Scanner(inputStream).useDelimiter("\\A");

String result = s.hasNext() ? s.next() : "";

4. 应用Stream Api (Java 8). 提示这种方式会将差别的换行符 (好比\r\n) 都替代为 \n.

String result = new BufferedReader(new InputStreamReader(inputStream))

  .lines().collect(Collectors.joining("\n"));

5. 应用parallel Stream Api (Java 8). 提示这种方式会将差别的换行符 (好比\r\n) 都替代为 \n.

String result = new BufferedReader(new InputStreamReader(inputStream))
  .lines().parallel().collect(Collectors.joining("\n"));

6. 应用InputStreamReader StringBuilder (JDK)

复制代码

final int bufferSize = 1024;

final char[] buffer = new char[bufferSize];

final StringBuilder out = new StringBuilder();

Reader in = new InputStreamReader(inputStream, "UTF-8");

for (; ; ) {

    int rsz = in.read(buffer, 0, buffer.length);

    if (rsz < 0)

        break;

    out.append(buffer, 0, rsz);

}

return out.toString();

复制代码

7. 应用StringWriter  IOUtils.copy (Apache Co妹妹ons)

StringWriter writer = new StringWriter();

IOUtils.copy(inputStream, writer, "UTF-8");

return writer.toString();

8. 应用ByteArrayOutputStream  inputStream.read (JDK)

复制代码

ByteArrayOutputStream result = new ByteArrayOutputStream();

byte[] buffer = new byte[1024];

int length;

while ((length = inputStream.read(buffer)) != -1) {

    result.write(buffer, 0, length);

}

return result.toString("UTF-8");

复制代码

9. 应用BufferedReader (JDK). 提示: 这种方式会将差别的换行符 (好比\r\n) 都替代为目前体系的换行符(比方windows下是"\r\n").

复制代码

String newLine = System.getProperty("line.separator");

BufferedReader reader = new BufferedReader(new InputStreamReader(inputStream));

StringBuilder result = new StringBuilder();

String line; boolean flag = false;

while ((line = reader.readLine()) != null) {

    result.append(flag? newLine: "").append(line);

    flag = true;

}

return result.toString();

复制代码

10. 应用BufferedInputStream  ByteArrayOutputStream (JDK)

复制代码

BufferedInputStream bis = new BufferedInputStream(inputStream);

ByteArrayOutputStream buf = new ByteArrayOutputStream();

int result = bis.read();

while(result != -1) {

    buf.write((byte) result);

    result = bis.read();

}

return buf.toString();

复制代码

11. 应用 inputStream.read()  StringBuilder (JDK). 提示这种方式处分Unicode时存在问题比方俄文仅在非Unicode字符串下事情正常.

复制代码

int ch;

StringBuilder sb = new StringBuilder();

while((ch = inputStream.read()) != -1)

    sb.append((char)ch);

reset();

return sb.toString();

复制代码


提示:
方式 4, 5  9 都存在替代换行符的问题方式11Unicode下不行正常事情.

性能测试
对于短字符串 (length = 175)的测试, url in github (mode = Average Time, system = Linux, score 1,343 is the best):

复制代码

              Benchmark                        Mode  Cnt   Score   Error  Units

8. ByteArrayOutputStream and read (JDK)        avgt   10   1,343 ± 0,028  us/op

6. InputStreamReader and StringBuilder (JDK)   avgt   10   6,980 ± 0,404  us/op

10.BufferedInputStream, ByteArrayOutputStream  avgt   10   7,437 ± 0,735  us/op

11.InputStream.read() and StringBuilder (JDK)  avgt   10   8,977 ± 0,328  us/op

7. StringWriter and IOUtils.copy (Apache)      avgt   10  10,613 ± 0,599  us/op

1. IOUtils.toString (Apache Utils)             avgt   10  10,605 ± 0,527  us/op

3. Scanner (JDK)                               avgt   10  12,083 ± 0,293  us/op

2. CharStreams (guava)                         avgt   10  12,999 ± 0,514  us/op

4. Stream Api (Java 8)                         avgt   10  15,811 ± 0,605  us/op

9. BufferedReader (JDK)                        avgt   10  16,038 ± 0,711  us/op

5. parallel Stream Api (Java 8)                avgt   10  21,544 ± 0,583  us/op

复制代码


对于长字符串的测试 (length = 50100), url in github (mode = Average Time, system = Linux, score 200,715 is the best):

复制代码

              Benchmark                        Mode  Cnt   Score        Error  Units

8. ByteArrayOutputStream and read (JDK)        avgt   10   200,715 ±   18,103  us/op

1. IOUtils.toString (Apache Utils)             avgt   10   300,019 ±    8,751  us/op

6. InputStreamReader and StringBuilder (JDK)   avgt   10   347,616 ±  130,348  us/op

7. StringWriter and IOUtils.copy (Apache)      avgt   10   352,791 ±  105,337  us/op

2. CharStreams (guava)                         avgt   10   420,137 ±   59,877  us/op

9. BufferedReader (JDK)                        avgt   10   632,028 ±   17,002  us/op

5. parallel Stream Api (Java 8)                avgt   10   662,999 ±   46,199  us/op

4. Stream Api (Java 8)                         avgt   10   701,269 ±   82,296  us/op

10.BufferedInputStream, ByteArrayOutputStream  avgt   10   740,837 ±    5,613  us/op

3. Scanner (JDK)                               avgt   10   751,417 ±   62,026  us/op

11.InputStream.read() and StringBuilder (JDK)  avgt   10  2919,350 ± 1101,942  us/op

复制代码

能够看出86是相对更好的选定

 

 

上篇:上一篇:Java Map 多见用法举例
下篇:下一篇:cycript装置遇到问题,装置失败