Bytes of a string in Java

A string is a list of characters (i.e. code points). The number of bytes taken to represent the string depends entirely on which encoding you use to turn it into bytes.

That said, you can turn the string into a byte array and then look at its size as follows:

// The input string for this test
final String string = "Hello World";

// Check length, in characters
System.out.println(string.length()); // prints "11"

// Check encoded sizes
final byte[] utf8Bytes = string.getBytes("UTF-8");
System.out.println(utf8Bytes.length); // prints "11"

final byte[] utf16Bytes= string.getBytes("UTF-16");
System.out.println(utf16Bytes.length); // prints "24"

final byte[] utf32Bytes = string.getBytes("UTF-32");
System.out.println(utf32Bytes.length); // prints "44"

final byte[] isoBytes = string.getBytes("ISO-8859-1");
System.out.println(isoBytes.length); // prints "11"

final byte[] winBytes = string.getBytes("CP1252");
System.out.println(winBytes.length); // prints "11"

So you see, even a simple "ASCII" string can have different number of bytes in its representation, depending which encoding is used. Use whichever character set you're interested in for your case, as the argument to getBytes(). And don't fall into the trap of assuming that UTF-8 represents every character as a single byte, as that's not true either:

final String interesting = "\uF93D\uF936\uF949\uF942"; // Chinese ideograms

// Check length, in characters
System.out.println(interesting.length()); // prints "4"

// Check encoded sizes
final byte[] utf8Bytes = interesting.getBytes("UTF-8");
System.out.println(utf8Bytes.length); // prints "12"

final byte[] utf16Bytes= interesting.getBytes("UTF-16");
System.out.println(utf16Bytes.length); // prints "10"

final byte[] utf32Bytes = interesting.getBytes("UTF-32");
System.out.println(utf32Bytes.length); // prints "16"

final byte[] isoBytes = interesting.getBytes("ISO-8859-1");
System.out.println(isoBytes.length); // prints "4" (probably encoded "????")

final byte[] winBytes = interesting.getBytes("CP1252");
System.out.println(winBytes.length); // prints "4" (probably encoded "????")

(Note that if you don't provide a character set argument, the platform's default character set is used. This might be useful in some contexts, but in general you should avoid depending on defaults, and always use an explicit character set when encoding/decoding is required.)

If you're running with 64-bit references:

sizeof(string) = 
8 + // object header used by the VM
8 + // 64-bit reference to char array (value)
8 + string.length() * 2 + // character array itself (object header + 16-bit chars)
4 + // offset integer
4 + // count integer
4 + // cached hash code

In other words:

sizeof(string) = 36 + string.length() * 2

On a 32-bit VM or a 64-bit VM with compressed OOPs (-XX:+UseCompressedOops), the references are 4 bytes. So the total would be:

sizeof(string) = 32 + string.length() * 2

This does not take into account the references to the string object.

The pedantic answer (though not necessarily the most useful one, depending on what you want to do with the result) is:

string.length() * 2

Java strings are physically stored in UTF-16BE encoding, which uses 2 bytes per code unit, and String.length() measures the length in UTF-16 code units, so this is equivalent to:

final byte[] utf16Bytes= string.getBytes("UTF-16BE");
System.out.println(utf16Bytes.length);

And this will tell you the size of the internal char array, in bytes.

Note: "UTF-16" will give a different result from "UTF-16BE" as the former encoding will insert a BOM, adding 2 bytes to the length of the array.

According to How to convert Strings to and from UTF8 byte arrays in Java:

String s = "some text here";
byte[] b = s.getBytes("UTF-8");
System.out.println(b.length);

Bytes of a string in Java

Related

Recent Posts