有一个开源项目可以做到这一点，甚至更多。

7.Bits().ToString();         // 7 b
8.Bits().ToString();         // 1 B
(.5).Kilobytes().Humanize();   // 512 B
(1000).Kilobytes().ToString(); // 1000 KB
(1024).Kilobytes().Humanize(); // 1 MB
(.5).Gigabytes().Humanize();   // 512 MB
(1024).Gigabytes().ToString(); // 1 TB

http://humanizr.net/#bytesize

https://github.com/MehdiK/Humanizer

string[] suffixes = { "B", "KB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB" };
int s = 0;
long size = fileInfo.Length;

while (size >= 1024)
{
    s++;
    size /= 1024;
}

string humanReadable = String.Format("{0} {1}", size, suffixes[s]);

这可能不是最有效或最优化的方法，但如果您不熟悉对数数学，它更容易阅读，并且对于大多数情况来说应该足够快。

string[] sizes = { "B", "KB", "MB", "GB", "TB" };
double len = new FileInfo(filename).Length;
int order = 0;
while (len >= 1024 && order < sizes.Length - 1) {
    order++;
    len = len/1024;
}

// Adjust the format string to your preferences. For example "{0:0.#}{1}" would
// show a single decimal place, and no space.
string result = String.Format("{0:0.##} {1}", len, sizes[order]);

下面是@ deepe1的BigInteger版本的答案，它绕过了long的大小限制(因此支持yottabyte和理论上的任何后面的限制):

public static string ToBytesString(this BigInteger byteCount, string format = "N3")
{
    string[] suf = { "B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "YiB" };
    if (byteCount.IsZero)
    {
        return $"{0.0.ToString(format)} {suf[0]}";
    }

    var abs = BigInteger.Abs(byteCount);
    var place = Convert.ToInt32(Math.Floor(BigInteger.Log(abs, 1024)));
    var pow = Math.Pow(1024, place);

    // since we need to do this with integer math, get the quotient and remainder
    var quotient = BigInteger.DivRem(abs, new BigInteger(pow), out var remainder);
    // convert the remainder to a ratio and add both back together as doubles
    var num = byteCount.Sign * (Math.Floor((double)quotient) + ((double)remainder / pow));

    return $"{num.ToString(format)} {suf[place]}";
}

又多了一种方法，不管怎样。我喜欢上面提到的@humbads优化解决方案，所以复制了原理，但我实现了一点不同。

我认为它是否应该是一个扩展方法是有争议的(因为不是所有的long都必须是字节大小)，但我喜欢它们，当我下次需要它时，我可以在某个地方找到它!

关于单位，我想我从来没有说过“Kibibyte”或“Mebibyte”，虽然我对这种强制而非进化的标准持怀疑态度，但我认为从长远来看，这将避免混淆。

public static class LongExtensions
{
    private static readonly long[] numberOfBytesInUnit;
    private static readonly Func<long, string>[] bytesToUnitConverters;

    static LongExtensions()
    {
        numberOfBytesInUnit = new long[6]    
        {
            1L << 10,    // Bytes in a Kibibyte
            1L << 20,    // Bytes in a Mebibyte
            1L << 30,    // Bytes in a Gibibyte
            1L << 40,    // Bytes in a Tebibyte
            1L << 50,    // Bytes in a Pebibyte
            1L << 60     // Bytes in a Exbibyte
        };

        // Shift the long (integer) down to 1024 times its number of units, convert to a double (real number), 
        // then divide to get the final number of units (units will be in the range 1 to 1023.999)
        Func<long, int, string> FormatAsProportionOfUnit = (bytes, shift) => (((double)(bytes >> shift)) / 1024).ToString("0.###");

        bytesToUnitConverters = new Func<long,string>[7]
        {
            bytes => bytes.ToString() + " B",
            bytes => FormatAsProportionOfUnit(bytes, 0) + " KiB",
            bytes => FormatAsProportionOfUnit(bytes, 10) + " MiB",
            bytes => FormatAsProportionOfUnit(bytes, 20) + " GiB",
            bytes => FormatAsProportionOfUnit(bytes, 30) + " TiB",
            bytes => FormatAsProportionOfUnit(bytes, 40) + " PiB",
            bytes => FormatAsProportionOfUnit(bytes, 50) + " EiB",
        };
    }

    public static string ToReadableByteSizeString(this long bytes)
    {
        if (bytes < 0)
            return "-" + Math.Abs(bytes).ToReadableByteSizeString();

        int counter = 0;
        while (counter < numberOfBytesInUnit.Length)
        {
            if (bytes < numberOfBytesInUnit[counter])
                return bytesToUnitConverters[counter](bytes);
            counter++;
        }
        return bytesToUnitConverters[counter](bytes);
    }
}

我的观点是:

千字节的前缀是kB(小写K) 由于这些函数用于表示目的，因此应该提供区域性，例如:CurrentCulture,“{0时。##}{1}"，文件大小，单位); 根据上下文，千字节可以是1000字节或1024字节。MB、GB等也是如此。