Reverse-Engineering Find Function

Consider the MATLAB relational operations: k = (a <= b); m = find(a <= b); where a is a real-valued row vector and b is either a real-valued row vector of the same length as a, or, a scalar.

The objective of this problem is to reverse-engineer MATLAB’s built in function find.

I have to write a MATLAB function, [m,k] = my_find(a,b); that implements the above relational operations using for-loops and if-statements only and does not use find.

The outputs k,m must be the same as those above, and k must be of type logical.

So far, I have:

function [m,k]=my_find(a,b)
for i=1:length(a)
    if a<=b
        k=1
        m=a(a<=b)
    else
        k=0
    end
end

However, when I run the script for the values

a=[1 2 0 -3 7]
b=[3 2 4 -1 7]

I get k=1 as a scalar and not as a vector of 1s. The result for m is correct though, m=[1 2 0 -3 7].

How can I get k to show me the results as a vector of values for all of the elements of a, instead of just the first element being repeated?

I have tried this with a different set of values, and k still sticks to the value of the first element.

Answer 1

[Walter Roberson]

if a(i) <= b(i)

but you are going to have to work more on returning the indices.

[Fangjun Jiang]

And you need to determine whether b is a scalar or a vector with the same size as a.

[Ammar]

I used a(i) <= b(i) and it properly returns the different values of k as they should be. However, if b is a scalar, then I get an error because there is only one element in b, and it cannot access an index beyond 1. Is it possible to use a for loop to indicate whether b is a scalar or a vector?

I have to use the same function for all conditions, so I don't think I am allowed to create a subfunction or separate function for when be is a scalar.

Thanks for all the help.

[Bjorn Gustavsson]

Well, you'll have to determine the sizes of both inputs, then make cases handling the different possibilities a scalar b 1-D array, a and b 1-D arrays and so on...