Picking data from a large data file

Hi there

I was wondering if there is an algorithm to pick random data from a large matrix (2000*3). Lets say I want output (with 3 columns) which have every 10th row of my matrix. I want to keep number of columns same. I tried this by a fellow but unable to get same number of columns. for example

    mat = rand(randi([3 3]), randi([3 3]))
    x_every5th=x(10:10:numel(x));
    x_every10th=x(20:20:numel(x));
    x_every20th=x(50:50:numel(x));

column 1 is lat 2 is lon and z is gravity value. Is it also possible to set a region by defining lat lon and pick every 20th and in same matrix set a region to pick every 5th point.

would be looking for people like image analyst to help with it :)

Thanks Also if I want to define change selection with in the matrix (lets say for values between