Separating rows by attribute into new matrices.

Hi there,

I'm hoping someone can help me with this seemingly simple objective. I have a bunch of CIELAB values stored in a 1493x3 matrix like this:

75.0762811831022 5.18089489085732 14.6428639876497

73.5614856255498 5.30582608119210 14.6982379202033

69.3445535098640 5.14055617598219 13.5092813281339

79.4934199520990 4.66394941442407 11.9900551014935

79.0393320098919 4.36754772094482 11.5537506643861

I'd like to have a bunch of new matrices that store each of these rows by their first value (L*-value), lets say new matrices which contain all three rows for values with L*79, L*75, L*73 ect.

So in this case I would end up with 4 new matrices for L*-values of 75, 73, 69 and 79 and each of these matrices contains the entire row.

Your help is much apreciated.

Regards

SH

Answer 1

[drummer]

Hi,

Would it work if instead of several matrices, you get a new matrix with a third index?

So each index of the third dimension is the a new row of your previous input matrix.

As in your example, you provide:

75.0762811831022 5.18089489085732 14.6428639876497

73.5614856255498 5.30582608119210 14.6982379202033

69.3445535098640 5.14055617598219 13.5092813281339

79.4934199520990 4.66394941442407 11.9900551014935

79.0393320098919 4.36754772094482 11.5537506643861

What I mean is, instead of

m1 = [75.07 , 5.18, 14.64];

m2 = [73.56, 5.30, 14,.69]; and so on, you'd rather have

newMatrix(1, :, 1) = [75.07 , 5.18, 14.64];

newMatrix(2, :, 2)= [[73.56, 5.30, 14,.69];

newMatrix = zeros(1,3,1493);
yourMatrix = rand(1493,3);
[rowLen, colLen] = size(yourMatrix);
for z = 1 : rowLen % creates indexes for the 3rd dimension of newMatrix
    for i = 1 : rowLen % reads your rows
        for j = 1 : colLen % reads your cols
            newMatrix(i,j,z) = squeeze(yourMatrix(i,j,:));        
        end
    end
end

Cheers

[Sascha Hein]

Hi @drummer - thanks for your help mate! Unfortunately that doesn't do it. I ran your sript and get a bunch of zeros. Remember that I have a really large array here (14983x3). The max value for L* is 85.5 and min for L* is 41.8 so I would need 41 arrays which contain all respective LAB values...

[drummer]

I tested it before posting. You should check the rows by doing this in the command line:

% checking the values in a single row
yourMatrix(1,:) % This shows the random values in the 1st row, from yourMatrix = rand(1493,3)

To see if newMatrix has the first row separated in newMatrix(1,:,1), type in the command window:

newMatrix(1,:,1) % This should show the same values as in the previous line.

Do not use ; so you can se the outputs.

These outputs should be the same, so you have your vectors separated in newMatrix

FYI: If you want to change the code to your input, adjust the matrix sizes accordingly. Not sure if it's a typo, but you previously wrote 1493, and now 14983. The code was just an example to your approach.

In another topic: In your sample matrix you had 5 rows, but you wanted 4. Which of both rows with 79 you should have?

If you want a smaller number of vectors, you must add a comparison step after the loop to find the values you want. Otherwise, # of 'separated' vectors = # of rows of your input matrix.

Cheers

[Sascha Hein]

Hi @drumner!

Thanks again for your help. I probably failed to explain properly what I'm trying to do - my appologies. So I have an array of 14983x3 (sorry for the previous typo) and these are L*a*b*-values. So for instance 75.07 , 5.18, 14.64 referrs to L*75.07 a*5.18 and b*14.64. This array that I have subsequently has 14983 L*values ranging from L*41 to L*84. What I need is 43 arrays from 41 all the way up to 84. Each of these should contain all L*a*b*-values for each of these 43 L*-groups.

For instance I can simply use:

x=LAB(LAB(:, 1)<61, :);

L60=x(x(:, 1)>60, :);

LAB is my 14983x3 array and these two lines will create a new array called "L60" which contains all L*a*b*-values with L*60:

60.8008798471768 4.85292890948374 11.1145675710373

60.7213587044256 10.2993584240051 17.7875693024272

60.5630488563623 14.5555978316287 38.0850677656340

60.4203311933741 6.82778759593458 12.8733422501348

60.3948687072869 6.49893315895765 14.4663606974790

60.3287778519643 7.02156654492907 12.2648732211747

60.3040045656528 8.68237804846326 19.4577309730639

60.2968866679374 11.0826197879960 19.5396305850156