I want to calculate velocity. How do I apply my code to all tables in all cells?

Hi,

I have a data set of distances (see "results_distances" attached) and want to calculate the velocity for each row in each table in each cell.

I have the following code:

results_velocity = cell(size(results_distances));

for i = 1:length(results_distances) % iterate over tables in cell array
    if ~isempty(results_distances{i}) % check if table is not empty
        table_data = results_distances{i};
        v_table = table_data; % initialize output table with the same size as input table

        % Calculate velocity for each numeric variable (column) in the table
        for var = 1:width(table_data)
            % Check if the variable is numeric
            if isnumeric(table_data{:, var})
                % Calculate velocity for each row
                for row = 2:size(table_data, 1)
                    % Subtract previous value from current value and divide by time_interval
                    v_table{row, var} = (table_data{row, var} - table_data{row-1, var}) / time_interval;
                end
            end
        end
        
        results_velocity{i} = v_table; % store velocity table in output cell array
    end
end

However, when I run my code, it seems to only apply the calculation to the first column of tables in the cell array (see "results_velocity"). What am I doing wrong?

Thanks!

Answer 1

[Voss]

load('results_distances.mat')

results_distances is of size 3x12

results_distances

results_distances = 3x12 cell array
  Columns 1 through 10

    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2154x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}

  Columns 11 through 12

    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}

so its length is 12. Recall that length() returns the size of the longest dimension.

length(results_distances)

ans = 12

Iterating from 1 to length(results_distances) is iterating over the first 12 elements of results_distances, which in this case is the first 4 columns (since results_distances has 3 rows).

A same-size example using a numeric matrix:

data = reshape(1:36,3,[])

data = 3x12

1 4 7 10 13 16 19 22 25 28 31 34 2 5 8 11 14 17 20 23 26 29 32 35 3 6 9 12 15 18 21 24 27 30 33 36

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

for i = 1:length(data)
    data(i)
end

ans = 1

ans = 2

ans = 3

ans = 4

ans = 5

ans = 6

ans = 7

ans = 8

ans = 9

ans = 10

ans = 11

ans = 12

In order to iterate over all elements of results_distances, use numel instead of length.

time_interval = 1;

results_velocity = cell(size(results_distances));

for i = 1:numel(results_distances) % iterate over tables in cell array
    if ~isempty(results_distances{i}) % check if table is not empty
        table_data = results_distances{i};
        v_table = table_data; % initialize output table with the same size as input table
        
        % Calculate velocity for each numeric variable (column) in the table
        for var = 1:width(table_data)
            % Check if the variable is numeric
            if isnumeric(table_data{:, var})
                % Calculate velocity for each row
                for row = 2:size(table_data, 1)
                    % Subtract previous value from current value and divide by time_interval
                    v_table{row, var} = (table_data{row, var} - table_data{row-1, var}) / time_interval;
                end
            end
        end
        
        results_velocity{i} = v_table; % store velocity table in output cell array
    end
end

results_velocity

results_velocity = 3x12 cell array
  Columns 1 through 10

    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2154x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}    {2155x2 table}

  Columns 11 through 12

    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}
    {2155x2 table}    {2155x2 table}

[Voss]

A side note: because you're initializing v_table = table_data; and then updating rows 2 through end of v_table, when the code is complete, the first row of v_table still contains the first row of distances from table_data.

[lil brain]

Hi @Voss, thanks a bunch. And yes, to your side note: that is because when calculating the velocity I need to subtract the prvious value from the current one which doesnt work for the first line I suppose. Is there any workaround you suggest? Deleting the first row from each table?

[Voss]

You're welcome!

Taking the difference between adjacent elements will produce an array of size one less than the original array, so, yeah, ending up with a table with one fewer row than the original table is what I would do.

You can use the diff function to do it. diff(_,1,1) is the first difference in the first dimension.

load('results_distances.mat')

time_interval = 1;

results_velocity = cell(size(results_distances));

for i = 1:numel(results_distances) % iterate over tables in cell array
    
    if isempty(results_distances{i}) % check if table is not empty
        continue % skip if table is empty
    end
    
    table_data = results_distances{i};
    
    idx = vartype('numeric');
    
    v = diff(table_data{:,idx},1,1) / time_interval;
    
    results_velocity{i} = array2table(v); % store velocity table in output cell array
    
end

results_velocity

results_velocity = 3x12 cell array
  Columns 1 through 10

    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2153x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}
    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}
    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}    {2154x2 table}

  Columns 11 through 12

    {2154x2 table}    {2154x2 table}
    {2154x2 table}    {2154x2 table}
    {2154x2 table}    {2154x2 table}

results_velocity{1}

ans = 2154x2 table
        v1             v2     
    ___________    ___________

     7.5049e-06     9.9013e-07
     1.5347e-05     7.4745e-06
    -2.0066e-05    -1.2233e-05
    -2.8578e-06     4.9597e-06
     2.0392e-06     2.6013e-06
     6.7307e-06    -2.8106e-06
     8.1676e-06     4.4089e-06
     4.2043e-06    -2.0911e-07
    -7.4221e-06     2.1905e-05
    -2.0846e-05     1.9876e-05
     6.5665e-06     1.9515e-05
     2.1943e-06     5.6662e-06
     1.0346e-05     1.6792e-05
    -4.9744e-06    -9.3965e-06
    -1.2482e-06    -7.6492e-06
    -4.7839e-06    -1.5141e-05

[lil brain]

Wow this is perfect. Thank you again! Incredibly helpful and saves me so much time!

[Voss]

You're welcome!