Pairs of elements that meet a condition?

Suppose all_val is an nx2 matrix of real numbers.

How to get a matrix cond_val with only the rows of all_val for which

fun(j1)*fun(j2) < 0 % for some scalar-input function fun.m

Where j1 is the element from column 1 and j2 is the corrosponding element from column 2 of the rows of all_val.

I've had a look at Find Array Elements That Meet a Condition but have had no luck.

Answer 1

[Guillaume]

If fun can only operate on scalar:

cond_val = all_val(prod(arrayfun(@fun, all_val), 2) < 0, :);

If fun can operate on whole arrays at once, then simply:

cond_val = all_val(prod(fun(all_val), 2) < 0, :);

Answer 2

[Massimo Zanetti]

A=rand(15,2);
f=@(x) x.^2; %or another function you define
C=cell2mat( arrayfun( @(k) f(A(k,1))*f(A(k,2))<0 , (1:15)' ) );

Answer 3

[Star Strider]

I have no idea what ‘fun’ does or what ‘all_vals’ contains, but you can do something like this to create a logical vector output:

all_vals = randi(9, 10, 2);                                         % Create Matrix
fun = @(x) x - 5;                                                   % Create Function
Out = (fun(all_vals(:,1)) .* fun(all_vals(:,2))) < 0;               % Logical Vector Output

Elements of ‘Out’ equal to 1 are true so the condition was met, 0 are false.