How to find minimum from a group and index at which minimum value is obtained ?

Hello all,

I have a matrix Engine_data_temp whose 5th column contains values of time (eg: 1;1;1;2;2...600) and 10th column contains values of fuel consumption. I am currently using code shown below to find unique value of time and minimum value of fuel consumed at that time:

(As suggested my one of the community member)

 [uniqueT,~,jk]=unique(Engine_data_temp(:,5));
 minFc=accumarray(jk,Engine_data_temp(:,10),[],@min);

I also need to find the index at which min values of fuel consumption are being selected at each unique time instances because I need this index to access other columns in matrix Engine_data_temp.

Thank you!

[Dhaval Lodaya]

I am looking for a method which does not use for loop for this as it slows down my code. I need to obtain index by vectorization or any other method that is fast.

Thanks in advance!

Answer 1

[Walter Roberson]

Create a small function

function r = min_and_idx(x)
 [minx, minidx] = min(x);
 r = {minx, minidx};

Then you use

   minFc = accumarray(jk, Engine_data_temp(:,10), [], @min_and_idx, {});

The result will be a cell array in which some entries might be empty, but if they are not empty then they will be a cell with minimum value and the index of the value relative to the locations that share the same jk value.

[Dhaval Lodaya]

As you said there might be some entries which might be empty and accumarray does not except NaN. I got this error:"The function 'min_and_idx' returned a non-scalar value" How do I deal with this? Thanks!

[Walter Roberson]

 function r = min_and_idx(x)
 [minx, minidx] = min(x);
 r = {[minx, minidx]};

[Dhaval Lodaya]

Thanks a lot! This works perfectly fine now and I liked the cell approach.