Byte to String to byte conversion puzzling error

Question

Could anyone please help spot the error? Here's the code:

   byte[] oriBytes = { 0xB0, 0x2D };                       // oriBytes -> 0xB0, 0x2D
   string oriInStr = Encoding.ASCII.GetString(oriBytes);   // oriInStr ->   "?-"
   oriBytes = Encoding.ASCII.GetBytes(oriInStr);           // oriBytes -> 0x3F, 0x2D

I can't get back the original bytes values of 0xB0, 0x2D.

Why are you decoding using ASCII? Is your input restricted to ASCII? — Piyush Parashar, Nov 24 '14 at 07:55
Use `Encoding.Default`, this implements the codepage as used in your OS — leppie, Nov 24 '14 at 07:57
I've just solved it. just use Encoding.Unicode instead of ASCII. I work with every bit of a byte for my electronic hardware, ARM and PIC.. — sam byte, Nov 24 '14 at 08:08
Just make sure you are aware of input encoding while decoding otherwise you will end up in similar errors. — Piyush Parashar, Nov 24 '14 at 08:23

score 8 · Answer 1 · edited Nov 24 '14 at 09:16

8

0xB0 is not a valid ASCII code. You can read here:

Any byte greater than hexadecimal 0x7F is decoded as the Unicode question mark ("?")

edited Nov 24 '14 at 09:16

Patrick Hofman

153,850
22
249
325

answered Nov 24 '14 at 07:51

Fratyx

5,717
1
12
22

score 4 · Answer 2 · answered Nov 24 '14 at 07:52

That is because appearantly .NET doesn't support the Extended ASCII table. Every value above 127 will produce ?, which is 63.

Hence, converting the ? back will result in 63.

When running the code with UTF8 encoding, you will see it goes to the extended page, since the newBytes in this sample returns 4 bytes instead of 2:

byte[] oriBytes = { 0xB0, 0x2D };
string oriInStr = Encoding.UTF8.GetString(oriBytes);
byte[] newBytes = Encoding.UTF8.GetBytes(oriInStr);

score 1 · Answer 3 · answered Nov 24 '14 at 07:51

1

In you byte[] 0xB0 changes to 176 and 0x2D change to 45. When converted from ASCII which has only 128 character 176 will give you ? (undefined) and 45 gives you -.

Try to debug the code and see what is happening.

answered Nov 24 '14 at 07:51

Piyush Parashar

866
8
20

score 0 · Answer 4 · answered Nov 24 '14 at 07:55

0

ahaah.. I got it! use Encoding.Unicode instead of ASCII. Beware guys... ;)

   byte[] oriBytes = { 0xB0, 0x2D };                         // oriBytes -> 0xB0, 0x2D
   string oriInStr = Encoding.Unicode.GetString(oriBytes);   // oriInStr ->   "?-"
   oriBytes = Encoding.Unicode.GetBytes(oriInStr);           // oriBytes -> 0xB0, 0x2D

answered Nov 24 '14 at 07:55

sam byte

697
9
11

this works yes if you dont actualy want the string to be represented in ascii. – Vajura Nov 24 '14 at 07:56
I tested it with a very long string (text), in fact copying half a page. Convert it to bytes array then back to string. I can get the original text... Just watch out about this simple mistake. It can lead to severe headache, at least for me... ;) – sam byte Nov 24 '14 at 08:03

score 0 · Answer 5 · answered Nov 24 '14 at 07:55

0

As others have mentioned .Net doesnt support extended ascii. To solve this problem you should cast the byte values to char which is essentialy int and it will map them correctly.

 byte[] oriBytes = { 0xB0, 0x2D };                      
 string oriInStr = "";
 for (int a = 0; a < oriBytes.Length; a++)
     oriInStr += (char)(oriBytes[a]);
 oriBytes = Encoding.ASCII.GetBytes(oriInStr);

answered Nov 24 '14 at 07:55

Vajura

1,112
7
16

Have you tested your code? It gives the very same results! 0x3F, 0x2D – sam byte Nov 25 '14 at 00:31
i did yes how are you printing it out? – Vajura Nov 25 '14 at 06:08

Byte to String to byte conversion puzzling error

5 Answers5