Produce equality matrix based on elements in vector.

Given two equally sized vectors A and B, is there any way to make a matrix C of 1's and zero's such that the kth row of C contains 1's wherever elements of B equal the kth element of A?

I can do it by looping through elements of A, but I want to know if there's a vectorised way of doing this to speed it up?

[Alexander Holmes]

Actually, think I figured out an answer. I can just use repmat to create a matrix of n copies of the vector, and use == with the vector to compare it that way.

[Guillaume]

"I can just use repmat"

You don't need repmat. Implicit expansion will take care of repeating the elements for you and will be faster. See my answer.

Answer 1

[Guillaume]

Trivially done.

%assuming A and B are both row vectors:
C = A.' == B;  

If they're both column vectors, transpose B instead.

[Alexander Holmes]

Ah yeah, that does it perfectly. That's a much better method.

Thank you!

Answer 2

[Fangjun Jiang]

Like this?

A=randi(10,5,1);
B=randi(10,5,1);
k=3;
C=(B==A(k))

Answer 3

[Bernd Wagner]

Does the Logical opperator C= A==B not do that work?

It compares values in Vector A and responds a logical value 1 if the Value is also on the same line in B. Hence your C vector will be a vector of 0 and 1 with 1 if A==B.

[Alexander Holmes]

That wasn't quite what I wanted. I wantred to compare the whole vector to each element of another vector, which would give me a matrix at the end rather than a vector. But the user above has given me what I needed.

Thank you though!

Answer 4

[darova]

Try bsxfun

    % make all combinations using bsxfun
C = bsxfun(@minus,b(:),a(:)');  % b - rows, a - columns
[i,j] = find(~C);           % find 'zero'
C1 = C*0;
C1(i,:) = 1;                % make entire row '1' if any element a==b