fwrite with fixed length of different data types

I have a know format file with different reading format. Example -

fread(fid,[1, 40], 'char')
fread(fid,[1, 4], 'long')
fread(fid,[1, 1], 'short')
% These are data blocks
fread(fid,[400 1],'double')
fread(fid,[400 1],'int')
fread(fid,[400 1],'float')

Now, i have to write the data in the same fashion

fwrite(fid,strwrite,'40*char')

something like that -

how can i pad and fwrite of other data structures

Answer 1

[Walter Roberson]

40*char is not a permitted fwrite precision .

The precision option is not intended to give information about how many elements are being written: instead it gives information about how each input element is to be represented as output. The number is determined by the size of the variable .

fwrite(fid,strwrite,'char')

If you need to pad then fwrite the pad values yourself .

fwrite(zeroes(1,7),'uint8')

for example .

It is not possible to fwrite two different datatypes in one call. (However memory mapping routines can input or output entire structure at a time .)

[Kanchibhotla Chandra Sekhar]

Suppose, i have a c-file which writes the file as -

char node[16];

Can i write the same file in matlab as -

fwrite(fid,pad(nodestr,16),'char');

Is this both represent the same ?

[Walter Roberson]

no, pad(nodestr,16) would pad adding 16 values along the first dimension leading to a 17 x something character array.

Also C char is not fixed width: it is the smallest integer datatype supported by the target that is at least 8 bits and is implementation dependent as to whether it is signed or not. You should avoid defining interfaces in terms of char . Traditional C up to C89 did not offer any explicitly 8 bit datatype .

MATLAB fwrite char precision uses as many bytes as needed to encode the unicode input as the destination encoding specified or defaulted at fopen time. char*1 is 8 bit and will saturate for input beyond 255. C's char is not "wide" character .

[Walter Roberson]

Some time ago I wrote a small function to fwrite() with padding or truncation to a given length. It is easy to do and makes it easier to write out fixed-width structures.