How to compress data into numerical intervals

I am trying to compress the water depth data from the given CTD.mat in the following manner.

pwd;

load('pwd\CTD.mat');
A = (1:22).';
for c = 1:21
    for cc(c) = CTD.Depth{1, c}(1:end)
        for A = 1:22
            if cc(c) < 0.5
                A = 1;
                CTD.Depth2{1, c}(A) = 0;
            elseif cc(c) < 1
                A = 2;
                CTD.Depth2{1, c}(A) = 0.5;
            elseif cc(c) < 1.5
                A = 3;
                CTD.Depth2{1, c}(A) = 1;
            elseif cc(c) < 2
                A = 4;
                CTD.Depth2{1, c}(A) = 1.5;
            elseif cc(c) < 2.5
                A = 5;
                CTD.Depth2{1, c}(A) = 2;
            elseif cc(c) < 3
                A = 6;
                CTD.Depth2{1, c}(A) = 2.5;
            elseif cc(c) < 3.5
                A = 7;
                CTD.Depth2{1, c}(A) = 3
            elseif cc(c) < 4
                A = 8;
                CTD.Depth2{1, c}(A) = 3.5
            elseif cc(c) < 4.5
                A = 9;
                CTD.Depth2{1, c}(A) = 4
            elseif cc(c) < 5
                A = 10;
                CTD.Depth2{1, c}(A) = 4.5
            elseif cc(c) < 5.5
                A = 11;
                CTD.Depth2{1, c}(A) = 5
            elseif cc(c) < 6
                A = 12;
                CTD.Depth2{1, c}(A) = 5.5
            elseif cc(c) < 6.5
                A = 13;
                CTD.Depth2{1, c}(A) = 6
            elseif cc(c) < 7
                A = 14;
                CTD.Depth2{1, c}(A) = 6.5
            elseif cc(c) < 7.5
                A = 15;
                CTD.Depth2{1, c}(A) = 7
            elseif cc(c) < 8
                A = 16;
                CTD.Depth2{1, c}(A) = 7.5
            elseif cc(c) < 8.5
                A = 17;
                CTD.Depth2{1, c}(A) = 8
            elseif cc(c) < 9
                A = 18;
                CTD.Depth2{1, c}(A) = 8.5
            elseif cc(c) < 9.5
                A = 19;
                CTD.Depth2{1, c}(A) = 9
            elseif cc(c) < 10
                A = 20;
                CTD.Depth2{1, c}(A) = 9.5
            elseif cc(c) < 10.5
                A = 21;
                CTD.Depth2{1, c}(A) = 10
            else
                A = 22;
                CTD.Depth2{1, c}(A) = 10.5
            end
        end
    end
end

Please tell me how to fix this code. Also, I would like to compress the salinity data in CTD.mat like the right part of the given image file, can you explain how?

[영택 조]

Additional question: 'If the water depth starts from 0.5 m or higher, make the 0 m part NaN and then compress the (depth) / (average salinity) values equally.' -> In this case, what command is needed to put NaN in place of 0 m? Should I use ?

Answer 1

[Dyuman Joshi]

Here's an approach -

%Load the data
load('CTD.mat');

arr = CTD.Depth;

n = numel(arr);
%Pre-allocate the variable in which the output is to be stored
CTD.Sal2 = cell(size(CTD.Sal));

for k=1:numel(arr)
    vec = arr{1,k};
    %Specify groups to sort data into
    %For each set of data maximum value might be differ, so define the bins accordingly
    m = ceil(max(vec)/0.5)*0.5;
    bins = 0:0.5:m;
    
    %Indices of data sorted according to groups defined
    idx=discretize(vec,bins);
    
    %Calculate the mean of the data corresponding to the group indices
    CTD.Sal2{1,k} = accumarray(idx, CTD.Sal{1,k}, [], @mean);
end

%Check the value for the 1st array as shown in the image
disp(CTD.Sal2{1,1})

    0.1303
    0.1037
    0.1421
    0.2306
    0.2723
    1.2362
    5.3560
    8.7040
   11.8039
   16.3450
   18.2737
   19.7480
   20.6050
   21.3095
   21.3767

[영택 조]

Thank you so much! However, I would like to ask you one more thing. In the depth section, if there is no part corresponding to the part where the original depth is changed to '0m' - where the original depth is less than 0.5m - I would like to insert NaN instead of omitting the 'corresponding to the changed depth 0m' position in Sal2. (Just added a second image that does this.)

[Dyuman Joshi]

As for the additional question, if there are no values for a part/group, you can choose which value to use instead.

Note that I have changed the variable here, from Sal to Temp, as per the image, so I can show the comparison.

%Load the data
load('CTD.mat');

arr = CTD.Depth;

n = numel(arr);

%Pre-allocate the variable in which the output is to be stored
CTD.Temp2 = cell(size(CTD.Temp));

for k=1:numel(arr)
    vec = arr{1,k};
    %Specify groups to sort data into
    %For each set of data maximum value might be differ, so define the bins accordingly
    m = ceil(max(vec)/0.5)*0.5;
    bins = 0:0.5:m;
    
    %Indices of data sorted according to groups defined
    idx=discretize(vec,bins);
    
    %Calculate the mean of the data corresponding to the group indices
    
    %% Fill with NaN for no values corresponding to any group
    %%                                                         vvv
    CTD.Temp2{1,k} = accumarray(idx, CTD.Temp{1,k}, [], @mean, NaN);  
end

%Check the value for the 5th array as shown in the image
disp(CTD.Temp2{1,5})

       NaN
   25.1637
   25.0067
   24.9443
   24.7835
   24.6407
   24.4175
   24.3163
   24.2953
   24.3550
   24.3454
   24.2655
   24.1381
   24.0896
   24.0622
   24.0068
   23.9789
   23.9674

[영택 조]

It's working! thank you! Can I ask one last question? To create data in CTD.Depth2 like on the left in the second image, can I just replace @mean with a different function handle? If so, what handle do I need?

[Dyuman Joshi]

For that, you can will have need some more manipulation.

Also, I have combined all the operations here -

load('CTD.mat');

arr = CTD.Depth;

n = numel(arr);

%Preallocation
CTD.Sal2 = cell(size(CTD.Sal));
CTD.Temp2 = cell(size(CTD.Temp));
CTD.Depth2 = cell(size(CTD.Depth));

%All commands in a single for loop
for k=1:numel(arr)
    vec = arr{1,k};
    
    m = ceil(max(vec)/0.5)*0.5;
    bins = (0:0.5:m)';
    
    idx=discretize(vec,bins);
    
    CTD.Sal2{1,k} = accumarray(idx, CTD.Sal{1,k}, [], @mean, NaN);
    CTD.Temp2{1,k} = accumarray(idx, CTD.Temp{1,k}, [], @mean, NaN);
    CTD.Depth2{1,k} = sign(accumarray(idx, 1, [], [], NaN)).*bins(1:end-1);
end

disp(CTD.Sal2{1,1})

    0.1303
    0.1037
    0.1421
    0.2306
    0.2723
    1.2362
    5.3560
    8.7040
   11.8039
   16.3450
   18.2737
   19.7480
   20.6050
   21.3095
   21.3767

disp(CTD.Temp2{1,5})

       NaN
   25.1637
   25.0067
   24.9443
   24.7835
   24.6407
   24.4175
   24.3163
   24.2953
   24.3550
   24.3454
   24.2655
   24.1381
   24.0896
   24.0622
   24.0068
   23.9789
   23.9674

disp(CTD.Depth2{1,5})

       NaN
    0.5000
    1.0000
    1.5000
    2.0000
    2.5000
    3.0000
    3.5000
    4.0000
    4.5000
    5.0000
    5.5000
    6.0000
    6.5000
    7.0000
    7.5000
    8.0000
    8.5000

[영택 조]

Thanks a lot! : - )