How can I easily compress and decompress Strings to/from byte arrays?

Question

I have some strings that are roughly 10K characters each. There is plenty of repetition in them. They are serialized JSON objects. I'd like to easily compress them into a byte array, and uncompress them from a byte array.

How can I most easily do this? I'm looking for methods so I can do the following:

String original = "....long string here with 10K characters...";
byte[] compressed = StringCompressor.compress(original);
String decompressed = StringCompressor.decompress(compressed);
assert(original.equals(decompressed);

I would use InflatorInputStream/DeflatorOutputStream with ByteArrayInput/OutputStream. — Peter Lawrey, May 13 '12 at 14:11
There's an easy-to-use 'zip' class out there... edit - it is here http://docs.oracle.com/javase/6/docs/api/java/util/zip/package-summary.html and seems to use the classes @peter mentioned. — Tony Ennis, May 13 '12 at 14:11
How about this? http://stackoverflow.com/questions/3649485/how-to-compress-a-string — Mikita Belahlazau, May 13 '12 at 14:13
just using `String` and `byte[]` this can't be more than a 10-15 line method, assuming the JSON is all ascii. If you have to do something utf-8 ish, add 10 more lines... — CodeClown42, May 13 '12 at 14:13

Peter Lawrey · Accepted Answer · 2012-05-13T14:31:43.427

28

You can try

enum StringCompressor {
    ;
    public static byte[] compress(String text) {
        ByteArrayOutputStream baos = new ByteArrayOutputStream();
        try {
            OutputStream out = new DeflaterOutputStream(baos);
            out.write(text.getBytes("UTF-8"));
            out.close();
        } catch (IOException e) {
            throw new AssertionError(e);
        }
        return baos.toByteArray();
    }

    public static String decompress(byte[] bytes) {
        InputStream in = new InflaterInputStream(new ByteArrayInputStream(bytes));
        ByteArrayOutputStream baos = new ByteArrayOutputStream();
        try {
            byte[] buffer = new byte[8192];
            int len;
            while((len = in.read(buffer))>0)
                baos.write(buffer, 0, len);
            return new String(baos.toByteArray(), "UTF-8");
        } catch (IOException e) {
            throw new AssertionError(e);
        }
    }
}

edited May 13 '12 at 14:31

answered May 13 '12 at 14:22

Peter Lawrey

525,659
79
751
1,130

2

Hello, why do you use `enum` instead of `class` here? Is it to prove a point? – black panda May 13 '12 at 14:34
8

Some people like to use enum classes as a way to implement singletons or static-only classes. Recommended by Joshua Bloch, writer of Effective Java. – Steve McLeod May 13 '12 at 14:36
Its to say there no instances of this class allowed. – Peter Lawrey May 13 '12 at 14:36
OK, you use it as a side effect. Thanks :) – black panda May 13 '12 at 14:38
Alternatively, you could declare (just) a private no-args constructor. – Stephen C May 13 '12 at 14:45
... and make the class final. – Peter Lawrey May 13 '12 at 15:20
1

Note to myself (and some others who might have wanted to achieve the same): you can't use that method to compress a `String` into another `String`, as the charset will mess up the byte content (see http://stackoverflow.com/questions/2544965/why-new-stringbytes-enc-getbytesenc-does-not-return-the-original-byte-array). – Matthieu Jul 18 '13 at 16:34
1

@Matthieu You can use ISO-8859-1 encoding to store bytes in a String without loss. – Peter Lawrey Jul 18 '13 at 18:03

score 3 · Answer 2 · answered Feb 17 '17 at 06:30

Peter Lawrey's answer can be improved a bit using this less complex code for the decompress function

    ByteArrayOutputStream baos = new ByteArrayOutputStream();
    try {
        OutputStream out = new InflaterOutputStream(baos);
        out.write(bytes);
        out.close();
        return new String(baos.toByteArray(), "UTF-8");
    } catch (IOException e) {
        throw new AssertionError(e);
    }

lithedream · Answer 3 · 2017-06-16T19:08:33.343

1

I made a library to solve the problem of compressing generic Strings (expecially short ones). It tries to compress the String using various algorithms (plain utf-8, 5bit encoding for latin letters, huffman encoding, gzip for long Strings) and chooses the one with the shortest result (in the worst case, it will choose the utf-8 encoding, so that you never risk to lose space).

I hope it may be useful, here's the link https://github.com/lithedream/lithestring

EDIT: I realized that your Strings are always "long", my library defaults on gzip for those sizes, I fear I cannot do better for you.

edited Jun 16 '17 at 19:08

answered Jun 16 '17 at 19:00

lithedream

11
4

Why a library if the standard API already solves the problem? – Lluis Martinez Mar 27 '19 at 13:27

How can I easily compress and decompress Strings to/from byte arrays?

3 Answers3

Linked