How to impove the performence of this code?

Hello,

I am doing a for-loop, but this computing cost is too huge for me. Therefore I am trying to find a shut-cut way to reduce the cost, I have no ideas how to do this vectorization.

Any suggestions are greatly appreciated.

Many thanks

Code:

--------------------------------------------------------

felspar(197992,128);

felspar(:,1:3)=alldatafel(:,1:3);

for j = 1:197992

    for i=1:337945   

        if strcmp(alldata(i,1),alldatafel(j,1)) && strcmp(alldata(i,2),alldatafel(j,2));

            felspar(j,4:128)=alldata(i,1:125);
        end                     
    end
end
xlswrite('match.xlsx',felspar)

Answer 1

[Hugo]

As far as I can see, there is a problem in the code. Wheneven the condition

strcmp(alldata(i,1),alldatafel(j,1)) && strcmp(alldata(i,2),alldatafel(j,2))

is TRUE, then

felspar(j,4:128)=alldata(i,1:125)

But that means that the value stored in felspar(j,4:128) for each j corresponds to the largest value of i for which the condition is fulfilled. Therefore, you can save time by searching i from the end, and breaking the loop as soon as the condition is fulfilled. The could should look like this:

felspar(197992,128);

felspar(:,1:3)=alldatafel(:,1:3);

for j = 1:197992

    for i=337945:-1:1   

        if strcmp(alldata(i,1),alldatafel(j,1)) &&  strcmp(alldata(i,2),alldatafel(j,2));

            felspar(j,4:128)=alldata(i,1:125);
            break;
  end                     
    end
  end

xlswrite('match.xlsx',felspar)

You could also save time by preserving in alldata and alldatafel only those items that are unique, using the unique function.

Hope this helps.