skip data when reading binary

I have a binary data file, and I am reading it as such.

fileID = fopen('2016.06.27_15.12.07_1.A.bin');
SDDdata = fread(fileID, 1E6, 'uint16' , 'ieee-le');
fclose all;

However, I want to skip every other data point, or every 3, etc. As the file is incredibly large and I want to to some testing with ignoring data. Looking up the fread command I can see there is a skip options, but I could not get it to work for this? It seemed to do nothing?

[James Tursa]

Please show the code you are using for skipping. Are you sure you are inputting the number of bytes to skip and not the number of values?

[Walter Roberson]

Good point about it being in bytes .

Answer 1

[Walter Roberson]

In R2016b I tested by writing out uint8(0:255) repeatedly, and then doing several tests

bytes = uint8(0:255);
fid = fopen('testfread.bin', 'w');
for K = 1 : 4096; fwrite(fid, bytes, 'uint8'); end
fclose(fid);

fid = fopen"testfread.bin', 'r');

skip = 0; frewind(fid); data0 = fread(fid, [1 100], 'uint8', skip, 'ieee-le')
skip = 1; frewind(fid); data1 = fread(fid, [1 100], 'uint8', skip, 'ieee-le')
skip = 2; frewind(fid); data2 = fread(fid, [1 100], 'uint8', skip, 'ieee-le')

fclose(fid)

The output was exactly what I expected: first consecutive bytes [0 1 2 3 ...] for skip 0, then every second byte [0 2 4 6 ...] for skip 1, then every third byte [0 3 6 9 ...] for skip 2. Skip seems to work correctly on my system.

[James Tursa]

Worked for me too, which is why I want OP to post the exact code being used.

[Image Analyst]

Why didn't you try it for uint16? for completeness, I do that below:

% Create sample data file.
bytes = uint16(0:255);
fid = fopen('testfread.bin', 'w');
for K = 1 : 4096; fwrite(fid, bytes, 'uint16'); end
fclose(fid);

% Read it back in.
fid = fopen('testfread.bin', 'r');

numbersToSkip = 0; 
bytesToSkip = numbersToSkip * 2; % Multiply by 2 for uint16, which is 2 bytes
frewind(fid); 
data0 = fread(fid, [1 100], 'uint16', bytesToSkip, 'ieee-le')

numbersToSkip = 1; 
bytesToSkip = numbersToSkip * 2; % Multiply by 2 for uint16, which is 2 bytes
frewind(fid); 
data1 = fread(fid, [1 100], 'uint16', bytesToSkip, 'ieee-le')

numbersToSkip = 2; 
bytesToSkip = numbersToSkip * 2; % Multiply by 2 for uint16, which is 2 bytes
frewind(fid); 
data2 = fread(fid, [1 100], 'uint16', bytesToSkip, 'ieee-le')

fclose(fid)

It's a little bit trickier since you have to multiply the number to skip by 2 since there are two bytes in a uint16 number.