您的位置:首页 > 编程语言 > Java开发

Java 字符串拼接 五种方法的性能比较分析 从执行100次到90万次

2016-09-30 16:29 621 查看
【请尊重原创版权,如需引用,请注明来源及地址】

> 字符串拼接一般使用“+”,但是“+”不能满足大批量数据的处理,Java中有以下五种方法处理字符串拼接,各有优缺点,程序开发应选择合适的方法实现。

1. 加号 “+”

2. String contact() 方法

3. StringUtils.join() 方法

4. StringBuffer append() 方法

5. StringBuilder append() 方法

> 经过简单的程序测试,从执行100次到90万次的时间开销如下表:



由此可以看出:

1. 方法1 加号 “+” 拼接 和 方法2 String contact() 方法 适用于小数据量的操作,代码简洁方便,加号“+” 更符合我们的编码和阅读习惯;

2. 方法3 StringUtils.join() 方法 适用于将ArrayList转换成字符串,就算90万条数据也只需68ms,可以省掉循环读取ArrayList的代码;

3. 方法4 StringBuffer append() 方法 和 方法5 StringBuilder append() 方法 其实他们的本质是一样的,都是继承自AbstractStringBuilder,效率最高,大批量的数据处理最好选择这两种方法。

4. 方法1 加号 “+” 拼接 和 方法2 String contact() 方法 的时间和空间成本都很高(分析在本文末尾),不能用来做批量数据的处理。

> 源代码,供参考

package cnblogs.twzheng.lab2;

/**
* @author Tan Wenzheng
*
*/
import java.util.ArrayList;
import java.util.List;

import org.apache.commons.lang3.StringUtils;

public class TestString {

private static final int max = 100;

public void testPlus() {
System.out.println(">>> testPlus() <<<");

String str = "";

long start = System.currentTimeMillis();

for (int i = 0; i < max; i++) {
str = str + "a";
}

long end = System.currentTimeMillis();

long cost = end - start;

System.out.println("   {str + \"a\"} cost=" + cost + " ms");
}

public void testConcat() {
System.out.println(">>> testConcat() <<<");

String str = "";

long start = System.currentTimeMillis();

for (int i = 0; i < max; i++) {
str = str.concat("a");
}

long end = System.currentTimeMillis();

long cost = end - start;

System.out.println("   {str.concat(\"a\")} cost=" + cost + " ms");
}

public void testJoin() {
System.out.println(">>> testJoin() <<<");

long start = System.currentTimeMillis();

List<String> list = new ArrayList<String>();

for (int i = 0; i < max; i++) {
list.add("a");
}

long end1 = System.currentTimeMillis();
long cost1 = end1 - start;

StringUtils.join(list, "");

long end = System.currentTimeMillis();
long cost = end - end1;

System.out.println("   {list.add(\"a\")} cost1=" + cost1 + " ms");
System.out.println("   {StringUtils.join(list, \"\")} cost=" + cost
+ " ms");
}

public void testStringBuffer() {
System.out.println(">>> testStringBuffer() <<<");

long start = System.currentTimeMillis();

StringBuffer strBuffer = new StringBuffer();

for (int i = 0; i < max; i++) {
strBuffer.append("a");
}
strBuffer.toString();

long end = System.currentTimeMillis();

long cost = end - start;

System.out.println("   {strBuffer.append(\"a\")} cost=" + cost + " ms");
}

public void testStringBuilder() {
System.out.println(">>> testStringBuilder() <<<");

long start = System.currentTimeMillis();

StringBuilder strBuilder = new StringBuilder();

for (int i = 0; i < max; i++) {
strBuilder.append("a");
}
strBuilder.toString();

long end = System.currentTimeMillis();

long cost = end - start;

System.out
.println("   {strBuilder.append(\"a\")} cost=" + cost + " ms");
}
}


> 测试结果:

1. 执行100次, private static final int max = 100;

>>> testPlus() <<<
{str + "a"} cost=0 ms
>>> testConcat() <<<
{str.concat("a")} cost=0 ms
>>> testJoin() <<<
{list.add("a")} cost1=0 ms
{StringUtils.join(list, "")} cost=20 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=0 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=0 ms


2. 执行1000次, private static final int max = 1000;

>>> testPlus() <<<
{str + "a"} cost=10 ms
>>> testConcat() <<<
{str.concat("a")} cost=0 ms
>>> testJoin() <<<
{list.add("a")} cost1=0 ms
{StringUtils.join(list, "")} cost=20 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=0 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=0 ms


3. 执行1万次, private static final int max = 10000;

>>> testPlus() <<<
{str + "a"} cost=150 ms
>>> testConcat() <<<
{str.concat("a")} cost=70 ms
>>> testJoin() <<<
{list.add("a")} cost1=0 ms
{StringUtils.join(list, "")} cost=30 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=0 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=0 ms


4. 执行10万次, private static final int max = 100000;

>>> testPlus() <<<
{str + "a"} cost=4198 ms
>>> testConcat() <<<
{str.concat("a")} cost=1862 ms
>>> testJoin() <<<
{list.add("a")} cost1=21 ms
{StringUtils.join(list, "")} cost=49 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=10 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=10 ms


5. 执行20万次, private static final int max = 200000;

>>> testPlus() <<<
{str + "a"} cost=17196 ms
>>> testConcat() <<<
{str.concat("a")} cost=7653 ms
>>> testJoin() <<<
{list.add("a")} cost1=20 ms
{StringUtils.join(list, "")} cost=51 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=20 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=16 ms


6. 执行50万次, private static final int max = 500000;

>>> testPlus() <<<
{str + "a"} cost=124693 ms
>>> testConcat() <<<
{str.concat("a")} cost=49439 ms
>>> testJoin() <<<
{list.add("a")} cost1=21 ms
{StringUtils.join(list, "")} cost=50 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=20 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=10 ms


7. 执行90万次, private static final int max = 900000;

>>> testPlus() <<<
{str + "a"} cost=456739 ms
>>> testConcat() <<<
{str.concat("a")} cost=186252 ms
>>> testJoin() <<<
{list.add("a")} cost1=20 ms
{StringUtils.join(list, "")} cost=68 ms
>>> testStringBuffer() <<<
{strBuffer.append("a")} cost=30 ms
>>> testStringBuilder() <<<
{strBuilder.append("a")} cost=24 ms


> 查看源代码,以及简单分析

String contact 和 StringBuffer,StringBuilder 的源代码都可以在Java库里找到,有空可以研究研究。

1. 其实每次调用contact()方法就是一次数组的拷贝,虽然在内存中是处理都是原子性操作,速度非常快,但是,最后的return语句会创建一个新String对象,限制了concat方法的速度。

public String concat(String str) {
int otherLen = str.length();
if (otherLen == 0) {
return this;
}
int len = value.length;
char buf[] = Arrays.copyOf(value, len + otherLen);
str.getChars(buf, len);
return new String(buf, true);
}


2. StringBuffer 和 StringBuilder 的append方法都继承自AbstractStringBuilder,整个逻辑都只做字符数组的加长,拷贝,到最后也不会创建新的String对象,所以速度很快,完成拼接处理后在程序中用strBuffer.toString()来得到最终的字符串。

/**
* Appends the specified string to this character sequence.
* <p>
* The characters of the {@code String} argument are appended, in
* order, increasing the length of this sequence by the length of the
* argument. If {@code str} is {@code null}, then the four
* characters {@code "null"} are appended.
* <p>
* Let <i>n</i> be the length of this character sequence just prior to
* execution of the {@code append} method. Then the character at
* index <i>k</i> in the new character sequence is equal to the character
* at index <i>k</i> in the old character sequence, if <i>k</i> is less
* than <i>n</i>; otherwise, it is equal to the character at index
* <i>k-n</i> in the argument {@code str}.
*
* @param   str   a string.
* @return  a reference to this object.
*/
public AbstractStringBuilder append(String str) {
if (str == null) str = "null";
int len = str.length();
ensureCapacityInternal(count + len);
str.getChars(0, len, value, count);
count += len;
return this;
}


/**
* This method has the same contract as ensureCapacity, but is
* never synchronized.
*/
private void ensureCapacityInternal(int minimumCapacity) {
// overflow-conscious code
if (minimumCapacity - value.length > 0)
expandCapacity(minimumCapacity);
}

/**
* This implements the expansion semantics of ensureCapacity with no
* size check or synchronization.
*/
void expandCapacity(int minimumCapacity) {
int newCapacity = value.length * 2 + 2;
if (newCapacity - minimumCapacity < 0)
newCapacity = minimumCapacity;
if (newCapacity < 0) {
if (minimumCapacity < 0) // overflow
throw new OutOfMemoryError();
newCapacity = Integer.MAX_VALUE;
}
value = Arrays.copyOf(value, newCapacity);
}


3. 字符串的加号“+” 方法, 虽然编译器对其做了优化,使用StringBuilder的append方法进行追加,但是每循环一次都会创建一个StringBuilder对象,且都会调用toString方法转换成字符串,所以开销很大。

  注:执行一次字符串“+”,相当于 str = new StringBuilder(str).append("a").toString();

4. 本文开头的地方统计了时间开销,根据上述分析再想想空间的开销。常说拿空间换时间,反过来是不是拿时间换到了空间呢,但是在这里,其实时间是消耗在了重复的不必要的工作上(生成新的对象,toString方法),所以对大批量数据做处理时,加号“+” 和 contact 方法绝对不能用,时间和空间成本都很高。

【请尊重原创版权,如需引用,请注明来源及地址】
内容来自用户分享和网络整理,不保证内容的准确性,如有侵权内容,可联系管理员处理 点击这里给我发消息
标签: