Converting string to byte[] creates zero character

问题:

In this convert function

public static byte[] GetBytes(string str)
{
    byte[] bytes = new byte[str.Length * sizeof(char)];
    System.Buffer.BlockCopy(str.ToCharArray(), 0, bytes, 0, bytes.Length);
    return bytes;
}

byte[] test = GetBytes("abc");

The resulting array contains zero character

test = [97, 0, 98, 0, 99, 0]

And when we convert byte[] back to string, the result is

string test = "a b c "

How do we make it so it doesn't create those zeroes

回答1:

First let's look at what your code does wrong. char is 16-bit (2 byte) in .NET framework. Which means when you write sizeof(char), it returns 2. str.Length is 1, so actually your code will be byte[] bytes = new byte[2] is the same byte[2]. So when you use Buffer.BlockCopy() method, you actually copy 2 bytes from a source array to a destination array. Which means your GetBytes() method returns bytes[0] = 32 and bytes[1] = 0 if your string is " ".

Try to use Encoding.ASCII.GetBytes() instead.

When overridden in a derived class, encodes all the characters in the specified string into a sequence of bytes.

const string input = "Soner Gonul";

byte[] array = Encoding.ASCII.GetBytes(input);

foreach ( byte element in array )
{
     Console.WriteLine("{0} = {1}", element, (char)element);
}

Output:

83 = S
111 = o
110 = n
101 = e
114 = r
32 =
71 = G
111 = o
110 = n
117 = u
108 = l

回答2:

Just to clear the confusion about your answer, char type in C# takes 2 bytes. So, string.toCharArray() returns an array in which each item takes 2 bytes of storage. While copying to byte array where each item takes 1 byte storage, there occurs a data loss. Hence the zeroes showing up in result.
As suggested, Encoding.ASCII.GetBytes is a safer option to use.

回答3:

In reality .net (at least for 4.0) automatically changes size of char when serialized with BinaryWriter

UTF-8 chars have variable length (might not be 1 byte), ASCII chars have 1 byte

'ē' = 2 bytes

'e' = 1 byte

It must be kept in mind when using

BinaryReader.ReadChars(stream)

In case of word "ēvalds" = 7 bytes size will be different than "evalds" = 6 bytes

回答4:

(97,0) is Unicode representation of 'a'. Unicode represents each character in two bytes. So you can not remove zeros. But you can change Encoding to ASCII. Try following for Converting string to byte[].

byte[] array = Encoding.ASCII.GetBytes(input);

回答5:

Try to specify Encoding explicitly. You can use next code to convert string to bytes with specified encoding

byte[] bytes = System.Text.Encoding.ASCII.GetBytes("abc");

if you print contents of bytes, you will get { 97, 98, 99 } which doesn't contain zeros, as in your example In your example default encoding using 16 bits per symbol. It can be observer by printing the results of

System.Text.Encoding.Unicode.GetBytes("abc"); // { 97, 0, 98, 0, 99, 0 }

Then while converting it back, you should select the appropriate encoding:

string str = System.Text.Encoding.ASCII.GetString(bytes);
Console.WriteLine (str);

Prints "abc" as you might expected