How to reduce execution time? "If" and "ismember" functions

I have a big code, and when analyze profiler results one of the consuming steps is the comparisson "if ii== jj" and "if nn==1"... Is there any other way to do this so I can reduce time?

if true
   if ii == jj
      ...
   else
       if nn==1
          ...
       else
          ...
       end
   end
end

The other function that is consuming a lot of time is "ismember". Also anyone knows a better way?

if true

    if ismember(nn,N1)
       ... 
    else
       ...
    end

end

Thanks!!

[Roger Wohlwend]

It surprises me that a simple command like if ii == nn is so time consuming. Are you sure that ii and nn are numbers and not vectors or matrices?

[Romain]

Can you specify the class, and size of ii, jj, nn and N1 ?

[Patty]

yes, and actually the numbers are not big.

ii=1:13 jj=1:13

nn=1:10

N1=1:5

Answer 1

[George Papazafeiropoulos]

To make ismember function faster, replace it with either ismembc or ismembc2 functions. ismembc returns an array of logical values while ismembc2 returns the index locations of the found members. More information about these can be found in the following:

http://undocumentedmatlab.com/blog/ismembc-undocumented-helper-function

https://www.mathworks.cn/matlabcentral/newsreader/view_thread/257728

Except for this, the "if ismember(nn,N1)" statement evaluates to true only if nn is a subset of N1. Since nn=1:10 and N1=1:5 it will always evaluate to false (the code after the above if statement will not be executed)

Since ii=1:13 and jj=1:13, the "if ii == jj" statement will evaluate always to true, preventing the next else section from execution.