我通常是这样做的:

public static String getFileSize(double size) {
    return _getFileSize(size,0,1024);
}

public static String _getFileSize(double size, int i, double base) {
    String units = " KMGTP";
    String unit = (i>0)?(""+units.charAt(i)).toUpperCase()+"i":"";
    if(size<base)
        return size +" "+unit.trim()+"B";
    else {
        size = Math.floor(size/base);
        return _getFileSize(size,++i,base);
    }
}

    public static String floatForm (double d)
    {
       return new DecimalFormat("#.##").format(d);
    }


    public static String bytesToHuman (long size)
    {
        long Kb = 1  * 1024;
        long Mb = Kb * 1024;
        long Gb = Mb * 1024;
        long Tb = Gb * 1024;
        long Pb = Tb * 1024;
        long Eb = Pb * 1024;

        if (size <  Kb)                 return floatForm(        size     ) + " byte";
        if (size >= Kb && size < Mb)    return floatForm((double)size / Kb) + " Kb";
        if (size >= Mb && size < Gb)    return floatForm((double)size / Mb) + " Mb";
        if (size >= Gb && size < Tb)    return floatForm((double)size / Gb) + " Gb";
        if (size >= Tb && size < Pb)    return floatForm((double)size / Tb) + " Tb";
        if (size >= Pb && size < Eb)    return floatForm((double)size / Pb) + " Pb";
        if (size >= Eb)                 return floatForm((double)size / Eb) + " Eb";

        return "???";
    }

有趣的事实:这里发布的原始代码片段是Stack Overflow上被复制最多的Java代码片段，它是有缺陷的。它被修好了，但却变得一团糟。 本文的完整故事:有史以来复制最多的堆栈溢出代码片段是有缺陷的!

来源:格式化字节大小到人类可读的格式|编程。指南

SI（1 k = 1，000）

public static String humanReadableByteCountSI(long bytes) {
    if (-1000 < bytes && bytes < 1000) {
        return bytes + " B";
    }
    CharacterIterator ci = new StringCharacterIterator("kMGTPE");
    while (bytes <= -999_950 || bytes >= 999_950) {
        bytes /= 1000;
        ci.next();
    }
    return String.format("%.1f %cB", bytes / 1000.0, ci.current());
}

二进制(1's = 1,024)

public static String humanReadableByteCountBin(long bytes) {
    long absB = bytes == Long.MIN_VALUE ? Long.MAX_VALUE : Math.abs(bytes);
    if (absB < 1024) {
        return bytes + " B";
    }
    long value = absB;
    CharacterIterator ci = new StringCharacterIterator("KMGTPE");
    for (int i = 40; i >= 0 && absB > 0xfffccccccccccccL >> i; i -= 10) {
        value >>= 10;
        ci.next();
    }
    value *= Long.signum(bytes);
    return String.format("%.1f %ciB", value / 1024.0, ci.current());
}

示例输出:

                             SI     BINARY

                  0:        0 B        0 B
                 27:       27 B       27 B
                999:      999 B      999 B
               1000:     1.0 kB     1000 B
               1023:     1.0 kB     1023 B
               1024:     1.0 kB    1.0 KiB
               1728:     1.7 kB    1.7 KiB
             110592:   110.6 kB  108.0 KiB
            7077888:     7.1 MB    6.8 MiB
          452984832:   453.0 MB  432.0 MiB
        28991029248:    29.0 GB   27.0 GiB
      1855425871872:     1.9 TB    1.7 TiB
9223372036854775807:     9.2 EB    8.0 EiB   (Long.MAX_VALUE)

试试JSR 363。它的单元扩展模块，如Unicode CLDR(在GitHub: uom-systems中)，为您完成所有这些。

你可以使用每个实现中包含的MetricPrefix或BinaryPrefix(与上面的一些例子相比)，如果你在印度或附近的国家生活和工作，IndianPrefix(也在uom-系统的公共模块中)允许你使用和格式化“千万字节”或“Lakh字节”。

private static final String[] Q = new String[]{"", "K", "M", "G", "T", "P", "E"};

public String getAsString(long bytes)
{
    for (int i = 6; i > 0; i--)
    {
        double step = Math.pow(1024, i);
        if (bytes > step) return String.format("%3.1f %s", bytes / step, Q[i]);
    }
    return Long.toString(bytes);
}

我最近问了同样的问题:

格式文件大小为MB, GB等。

虽然没有开箱即用的答案，但我可以接受这个解决方案:

private static final long K = 1024;
private static final long M = K * K;
private static final long G = M * K;
private static final long T = G * K;

public static String convertToStringRepresentation(final long value){
    final long[] dividers = new long[] { T, G, M, K, 1 };
    final String[] units = new String[] { "TB", "GB", "MB", "KB", "B" };
    if(value < 1)
        throw new IllegalArgumentException("Invalid file size: " + value);
    String result = null;
    for(int i = 0; i < dividers.length; i++){
        final long divider = dividers[i];
        if(value >= divider){
            result = format(value, divider, units[i]);
            break;
        }
    }
    return result;
}

private static String format(final long value,
    final long divider,
    final String unit){
    final double result =
        divider > 1 ? (double) value / (double) divider : (double) value;
    return new DecimalFormat("#,##0.#").format(result) + " " + unit;
}

测试代码:

public static void main(final String[] args){
    final long[] l = new long[] { 1l, 4343l, 43434334l, 3563543743l };
    for(final long ll : l){
        System.out.println(convertToStringRepresentation(ll));
    }
}

输出(在我的德语地区):

1 B
4,2 KB
41,4 MB
3,3 GB

我已经打开了一个问题，要求谷歌番石榴的这个功能。也许有人愿意支持它。