when I use fwrite to save a string with various ASCII characters, the resulting texfile changes some characters

Im trying to make a LZ78 coder, and when I want to save a string of characters using fwrite, the resulting .txt file is slightly different from the original string.

For instance, the string token_fin contains, as the first 10 characters, the following characters (here with their ASCII value):

24   136     2   217    88    52   134   122   216    24

When I use the next lines:

fwrite(write_file, token_fin, '*char');
fclose(write_file);

Then, when I use the next lines to open the file created previosly

read_file = fopen('LZ78_Output.txt','r');
labuffer = fread(read_file, '*char');

the first 10 characters are theese:

24    26     2   217    88    52    26   122   216    24

As you can see, characters 136 and 134 are replaced with 26.

Im not sure why this is happening, if anyone can help me I would really apreciate it.

Answer 1

[Cris LaPierre]

I think the issue is that you are using fwrite with *char. According to the precision table, this allocates 8 bits for the encoding. 8 bits allows for numbers from 0-127. Any value greater than that gets wrapped around in some way I have figured out yet. You need to set an appropriate machine format and enoding in fopen. It depends on where you are and your machine what you use. I used (..., 'n', 'ISO-8859-1').

Here is an approach that works for me.

token_fin = [24   136     2   217    88    52   134   122   216    24];
write_file = 'LZ78_Output.txt';
fid = fopen(write_file,'w+','n', 'ISO-8859-1');
fwrite(fid, token_fin, 'char');
fclose(fid);

read_file = fopen('LZ78_Output.txt','r','n', 'ISO-8859-1');
labuffer = fread(read_file, 'char');
double(labuffer)'

ans = 1×10    
    24   136     2   217    88    52   134   122   216    24

There are a couple other MATLAB Answer posts about this. See here and here for more details.

[Ricardo Fuertes]

Thank you so much! this worked. I now have the correct ASCII values. :)