最佳实践/性能: 将 StringBuilder.append 与 String.concat 混合

我试图了解什么是最佳实践,以及为什么要为不同的情况连接字符串文字和变量。例如,如果我有这样的代码

StringBuilder sb = new StringBuilder("AAAAAAAAAAAAA")
.append(B_String).append("CCCCCCCCCCC").append(D_String)
.append("EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE")
.append("FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF");

是这样做的吗?从 这篇文章开始,我注意到 String 上的 +操作符创建了一个 StringBuilder 的新实例,连接操作数,并返回一个 String 转换,这看起来比仅仅调用 .append()要复杂得多; 所以如果这是真的,那就不可能了。那 String.concat()呢?对每个连接使用 .append()合适吗?或者只是为了变量,而且字面值可以附加到 .concat()

StringBuilder sb = new StringBuilder("AAAAAAAAAAAAA")
.append(B_String.concat("CCCCCCCCCCC")).append(D_String
.concat("EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE")
.concat("FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF"));

处理这些情况的最佳实践和性能的一般规则是什么?我对 +的假设是正确的吗? 它真的不应该被使用吗?

89582 次浏览

The compilier optimize the + concatenation.

So

int a = 1;
String s = "Hello " + a;

is transformed into

new StringBuilder().append("Hello ").append(1).toString();

There an excellent topic here explaining why you should use the + operator.

You should always use append.

concat create a new String so it's pretty like + I think.

If you concat or use + with 2 final String the JVM can make optimisation so it's the same as doing append in this case.

+ operator

String s = s1 + s2

Behind the scenes this is translated to:

String s = new StringBuilder(s1).append(s2).toString();

Imagine how much extra work it adds if you have s1 + s2 here:

stringBuilder.append(s1 + s2)

instead of:

stringBuilder.append(s1).append(s2)

Multiple strings with +

Worth to note that:

String s = s1 + s2 + s3 + ... +sN

is translated to:

String s = new StringBuilder(s1).append(s2).append(s3)...apend(sN).toString();

concat()

String s = s1.concat(s2);

String creates char[] array that can fit both s1 and s2. Copies s1 and s2 contents to this new array. Actually requires less work then + operator.

StringBuilder.append()

Maintains an internal char[] array that grows when needed. No extra char[] is created if the internal one is sufficiently big.

stringBuilder.append(s1.concat(s2))

is also performing poorly because s1.concat(s2) creates an extra char[] array and copies s1 and s2 to it just to copy that new array contents to internal StringBuilder char[].

That being said you should use append() all the time and append raw strings (your first code snippet is correct).

If you concat exactly two Strings use String.concat (creates a new String by creating a new char-array that fits both Strings and copys both Strings' char arrays into it).

If you concat multiple (more than two) Strings in one line, use + or StringBuilder.append, it doesn't matter, since the compiler converts + to StringBuilder.append. This is good for multiple Strings because it maintains one char array that grows as needed.

If you concat multiple Strings over multiple lines create one StringBuilder and use the append-method. In the end when you are done appending Strings to the StringBuilder use it's .toString()-method to create a String out of it. For concatting in multiple lines this is faster than the second method, since the second method would create a new StringBuilder on each line, append the Strings and then cast back to String, while the third method only uses one StringBuilder for the whole thing.

Optimization is done automatically by the compiler.

The Java2 compiler will automatically convert the following:

String s = s1 + s2;

to

String s = (new StringBuffer()).append(s1).append(s2).toString();

Taken straight from the Java Best Practices on Oracles website.

Use + operator is best practice, it is also simple and readable.

The Java language provides special support for the string concatenation operator ( + ), and for conversion of other objects to strings. String concatenation is implemented through the StringBuilder(or StringBuffer) class and its append method.

Offical document: https://docs.oracle.com/javase/8/docs/api/java/lang/String.html

in byte code level there is not different and we are not compromising effeciancy there. In case of executing byte code level, it must go through non-inline operator overloading method for + by calling append. then in assembly language level (Java is written in C and C produces assemblies similar to assembly, there will be extra register call to store + method call in the stack and there will additional push. (in reality, cross-compiler might optimise + operator call, in that case making it no difference with efficiancy.)

It is a good practice to have one way to increase the readability. :)

I personally prefer the Strings.format(), simple easy to read one-line string formatter.

String b = "B value";
String d = "D value";
String fullString = String.format("A %s C %s E F", b, d);
// Output: A B value C D value E F

All answers are pretty good and explanatory. But I felt exploration around other string concatenation techniques would also help like - Guava Joiner, Streams, String.format etc.

For complete details over performance of each concatenation technique java-string-concatenation-which-way-is-best.

In Brief, concatenation performance varies with no. of strings to concatenate. For example - to concatenate 1-10 strings, these techniques works best - StringBuilder, StringBuffer and Plus Operator. And to concatenate 100s of strings - Guava Joiner, apache's stringsUtils library also works great.

Please go through the above blog. It really explains performance efficiency very well.

Thanks.