What is wrong with this

I just want to create a variable to hold 5 different sets of values...

Mass_red

Any ideas??

load remove_outlier.mat
LMT = [20 25 28 32 36];
HMT = [40 40 40 40 39.5];
for k=1:5
  i=1;
  while(Mass(i) < LMT(k))
      i=i+1;
  end

    j=1;
    while(Mass(j) < HMT(k))
        j=j+1;
    end

    Mass_red(k) = Mass(i:j);
end

Thanks

Note: Mass is loaded in to the workspace. It is a 3201 x 1 double. Basically, it is a vector with sequential values from 15 to 47. I suppose you could say it was 15:0.01:47

Answer 1

[Andrei Bobrov]

z = bsxfun(@times,Mass,bsxfun(@ge,Mass,LMT)&bsxfun(@lt,Mass,HMT));
Mass_red = z(any(z,2),:)';

OR [EDIT]

t = bsxfun(@ge,Mass,LMT)&bsxfun(@le,Mass,HMT);
z = bsxfun(@times,Mass,t)
Mass_red = arrayfun(@(ii)z(t(:,ii),ii),(1:size(z,2)),'un',0);

[S]

I have no idea what that is but I get an error when I copy it to the command window:

??? Error using ==> bsxfun

Non-singleton dimensions of the two input arrays must

match each other.

[S]

Actually it works. But it is just padding them all up with zeros to make them the same...

So yes, let me try again. Pretty good. Still dont know what it means :-)

[S]

I suppose it works but its not what I need. I dont want all these zeros padding up the vector.

Really I want like an array of vectors.

[Sean de Wolski]

triple bsxfuns!

+1

Answer 2

[Thomas]

To begin with:: you do not define Mass but you still use it to check its value in a while loop.

what is Mass?

EDIT

I think This is what you need

Mass=15:0.01:47;

LMT = [20 25 28 32 36];
HMT = [40 40 40 40 39.5];
for k=1:5
  i=1

    if (Mass(i) < LMT(k))
        i=i+1
    end

      j=1
      if (Mass(j) < HMT(k))
          j=j+1
      end

      Mass_red(k,:) = Mass(i:j);
  end

[S]

Sorry. My bad.

I missed out a line.

I have amended the original post.

Thanks

[Sean de Wolski]

abbreviation for the great state of Massachusetts?

[S]

??? Subscripted assignment dimension mismatch.

Mass_red(k,:) = Mass(i:j);

[Thomas]

above code works just fine for me.. maybe you have Mass '

after import try Mass=Mass'

[S]

thanks but no luck.

When I run in debug, it works for the first iteration (k=1), and create Mass_red 1 x 2001 double. But then on the next iteration (k=2), it crashes after implementing the Mass_red(k,:) line of code and gives the error (previous comment).

[Thomas]

what output does size(Mass) give you? For me Mass is 1x3201 and output Mass_red=5x2001. Also try a 'clear variables' at the beginning before you import your data.. ( though most people hate clearing..)

[S]

Thanks but neither of the suggestions work.

size(Mass) gives: 3201 1

[Thomas]

Does copy pasting the above code directly work for you? It does for me..

[S]

haha yes!

Why is that?? I just copied it in to the command window and it worked.

[S]

Saying that though... your code is wrong... Mass(i:j) length should change each time since we are cutting it off at different values LMT and HMT

[S]

It is wrong because... Mass(i) will always be less than LMT(k) and Mass (j) will always be less than HMT(k). Therefore i and j will increment only by 1 for each k.

Answer 3

[Walter Roberson]

i:j is going to have varying range depending on the data values, so Mass(i:j) will be varying sizes. You are trying to store each iteration into a numeric array, but the iterations are not of constant size. That is going to be a problem: numeric arrays cannot have rows or columns of varying size. You are going to have to use a cell array:

Mass_red{k} = Mass(i:j);

Notice the {} instead of ()

[S]

Thanks for advice but:

??? Cell contents assignment to a non-cell array

object.

Mass_red{k} = Mass(i:j);

[S]

Any advance on this...

Clearly cells are the way forward but I cant assign a non-cell array object to a cell... How can I do this???

[Walter Roberson]

That error implies that you initialized Mass_red in some code you did not show us. You need to initialize it differently for cell arrays.

Answer 4

[Image Analyst]

Instead of all that looping over i and j, why not just do

LMTIndex = find(Mass >= LMT(k), 1, 'first');
HMTIndex = find(Mass >= HMT(k), 1, 'first');
startingIndex = min([LMTIndex HMTIndex]);
endingIndex = max([LMTIndex HMTIndex]);
Mass_red{k} = Mass(startingIndex : endingIndex)