The problem with the function 'all'

Hello,

I need to compare two data sets, i.e.

a = [6 11 9; 5 4 3]; b = [1 2 3];

all(a>b); % The problem is that 'a' contains multiple rows, while 'b' - only a single row.

Therefore, the is an error: Error using > Matrix dimensions must agree.

But I'm expecting the following:

ans =

1

1

How can solve this?

Answer 1

[Sven]

I think you want to compare a(1,:) to b, and a(2,:) to b. So try this using bsxfun:

all(bsxfun(@gt, a, b),2)

Does that give what you expected?

Answer 2

[Walter Roberson]

all(a > repmat(b,size(a,1),1),2)

Answer 3

[Jan]

All values are smaller, if the smallest is smaller:

a = [6 11 9; 5 4 3];
b = [1 2 3];
min(a, [], 1) > b

For a=rand(1000, 1000), b = rand(1, 1000) this 30% faster than the BSXFUN approach, I assume because it needs less temporary memory.

[Sven]

I was thinking something like this might be more efficient. I think the answer that mimics:

all(bsxfun(@gt, a, b),2)

would if fact be:

min(a,[],2) > max(b)