Using find command on each term in a vector

I have two vectors that I am trying to use in my calculation.

One is a list of coordinates. A=-10:10

The other is a list of random numbers that may or may not be between these values. B=22*rand(1,30)-11

What I am trying to do is find the first value in A that is greater than each term in B. If B(1)=6.4, then I want the find command to give me the index in A that corresponds to 7. I want this value for each term in the B array, even if it’s the “empty 1-by-0 matrix.”

The syntax for that on each individual term in B is easy enough, but my question is: Is there a way to use the find command on the entire vector B at the same time and store that result in a third vector, or do I need to use using a for C=1:size(B,2) loop?

Answer 1

[Honglei Chen]

Not really a syntax for find, but take advantage of arrayfun

a = -10:10;
b = 22*rand(1,30)-11;
c = arrayfun(@(x) find(a>x,1), b, 'UniformOutput', false)

[James]

This worked with one caveat. If there's a value in B that's not between the high and low values of A, you get an error at that point with the command as shown. There's a flag 'Uniform Output' that you'll need to set, since not all of your output will be of the same type.

arrayfun(@(x) find(a>x,1), b,'Uniform Output',0)

I'd never heard of the arrayfun (or bsxfun for that matter) before. Thanks for the help.

[Sean de Wolski]

Welcome to the bsxfun club!

[Honglei Chen]

@James, you are right, I forgot that part, I've updated the answer for future reference.

Answer 2

[Andrei Bobrov]

A=-10:10
B=22*rand(1,30)-11
out = find(all(bsxfun(@gt,A.',B),2),1,'first')

[Honglei Chen]

Hi Andrei, it is an interesting thought to use bsxfun, but it doesn't seem to give the right answer. Am I missing something?

[James]

I tried this, and got an empty matrix. The bsxfun(@gt,A.',B) part worked fine, but the all(bsxfun(@gt,A.',B),2) gave me a column of zeroes. Since there were no ones in the matrix, the find part of this command gave me an empty mtarix.

Answer 3

[Sean de Wolski]

I don't see why you need to compare every value. Only compare values to the max of B since that is a requirement.

A=-10:10;
B=22*rand(1,30)-11;
idx = find(A>max(B),1,'first');
A(idx)

[James]

If I'm understanding what your command does (and I may be wrong on this, since I'm a relative MATLAB newbie), it compares the max value of B to each value in A. If the maximum value of B is greater than everything in A (which is quite possible) then you'll end up with idx being an empty matrix because no values will meet the criteria.

If the first four values of B are 6.4, -10.2, -1.3, 10.6, what I would like are the indices in A that correspond to 7, -10, -1 and 11 repsectively. Since there is no 11 in A, the output value for that index would end up being an empty matrix, so the output for these values would be [18 1 10 [] ]

(As an aside, I just used these values of A and B for simplicity. Sadly, my real data set is not this neat so I can't use floor and ceil.)