Sort one set of data to correspond to another.

Say I have an ordered set of data.

 a = [100,200,300,400,500];

And say I have another set of data,

 b = [300,200,500,400,100]

I m trying to find the index where a is sorted to b. I could use a nested for loop of course but is there a better way to get the index shown below?

 c = [5,2,1,4,3]

Answer 1

[Thorsten]

If b is an unsorted version of a, i.e., all elements in b occur once and only once in a, you can use

[~, idx] = sort(b);

In the more general case where b can have some elements of a, and elements can occur more than once, use

for i=1:numel(b), idx(i) = find(ismember(a, b(i))); end

Instead of the for loop, you can also use

idx = arrayfun(@(x) find(ismember(a,x)), b);

[Ronan]

So I probably asked the question the wrong way. I m looking for the index of one set of data that corresponds to another eg.

 x = [350,420,245,100]
 k = [420,100,350,245]

I d like to find where the index of x is in k. So the answer should be

 z = [3,1,4,2]

[Thorsten]

This is handled in the second part of my answer dealing with "the more general case"

z = arrayfun(@(i) find(ismember(k,i)), x)

[Ronan]

Thank you very much.

[Guillaume]

z = arrayfun(@(i) find(ismember(k,i)), x)

Is actually not going to work in the general case, because find may return none or several indices, which would then require a 'UniformOutput', false to arrayfun.

If you assume that find is always going to return one and only one value. Then the second output of ismember is a much more efficient way (no loop) to obtain the same result.

Answer 2

[Stephen23]

>> a = [100,200,300,400,500];
>> b = [300,200,500,400,100];
>> [~,Xa] = sort(a);
>> [~,Xb] = sort(b);
>> Xa(Xb)
ans =
   5   2   1   4   3

Answer 3

[Guillaume]

Following your comment to Thorsten's answer, use the 2nd return value of ismember:

x = [350,420,245,100]
k = [420,100,350,245]

z = ismember(x, k)