float array to binary string and vice-versa?

Hello, In order to exchange information with an external application, I need to "pack" an array of numbers into a string, and vice-versa. How can I achieve this? I need to do this with 32bit floats and 32bit unsigned integers.

The external application sends the data like following:

float myFloatArray[4];
char* dataToSend=(char*)myFloatArray;
sendData(dataToSend,4*4); // 4*4 is the size of the data

Data is packed as little Endian.

Thanks for any insight!

[James Tursa]

Are you using loadlibrary for this?

Answer 1

[José-Luis]

I don't exactly understand what you are trying to achieve in Matlab. The code you show is C/C++. Also, there are no pointers in Matlab (not explicitely at least). You could save your floats as a binary file, using little endian byte order. There is no need to typecast:

your_float = single(rand(1,4));
%Saving it as binary stream
fid = fopen('myBin.bin','w','l'); %little endian
fwrite(fid,your_float,'single');
fclose(fid);

The you could read that file in your C/C++ code. Or do you actually mean that you want a string of zeros and ones? That is a different thing.

Answer 2

[Marc]

Hello and thank you for your reply.

Actually I have to communicate with a C library. The C library is providing me with a string that contains coded (or packed) floats.

The C library puts the floats into a string according to the little-Endian standard. One 32bit float can be represented as 4 bytes (or chars in the string). It is like sending to Matlab the memory print of that float.

So, as an example, when the C library sends 3 floats, Matlab will receive 3*4 chars, or a string with 12 chars. For example on the C library side:

float my3Floats[3]={0.0f,1.0f,12345678.0f};
char* myStringToSend=(char*)my3Floats;

The 12 chars or byte values of "myStringToSend" contain:

0,0,0,0,0,0,128,63,78,97,60,75

My Matlab program will receive the above string that contains 12 bytes. From those 12 chars I need to rebuild the 3 floats. How do I do that in Matlab, without going through a file?

Basically:

x 32-bit floats in C are put into a x*4 char long string that is sent to Matlab. Matlab takes that string, and extracts 3 float values

Thanks

[José-Luis]

In that case maybe a mex file is what you want. You can create a mex function that accepts a pointer as input and returns the floats as output. You would need to do the relevant typecasting in C.

 doc mex