How to return result as NA rather than invalid line

Question

Ismat Mohd Sulaiman 2021-7-26

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/886384-how-to-return-result-as-na-rather-than-invalid-line

评论： Dave 2021-7-27

Given the simple data below, I want to extract the result from the previous rows (L1 = rows-1, and L2 = rows-2). But L2 does not exist and MATLAB return as invalid and will not proceed. How to I return L2 = NA and allow the script to proceed?

data = [1 2 3; 4 5 6; 7 8 9; 10 11 12]
[nrows, ncols] = size(data)    %determine the number of rows and columns 
for rows = 1:nrows      %start with row 1 to nrows (number of rows in the table)
    if data(rows,2) == 5        %and if that rows is equal to what we want
        disp("Row " + rows + " has number 5")
        L1 = data(rows-1, 2)
        L2 = data(rows-2, 2)  
        % The L2 row doesn't exist. Running this will return invalid. How to show L2 = NA and not be invalid
    else
        disp("Row " + rows + " has no number 5")
    end
end

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Dave B 2021-7-26

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/886384-how-to-return-result-as-na-rather-than-invalid-line#answer_754284

在 MATLAB Online 中打开

How about something like this?

if rows-2 > 0 && rows-2 <= nrows
    L2 = data(rows-2, 2);
else
    L2 = nan;
end

2 个评论
显示无隐藏无

Ismat Mohd Sulaiman 2021-7-27

Yup. Thanks. After I posted, I simply added another if else to the script, and it worked. 😊

Now I am thinking how to store all the results for L1 and L2 in an array 😁

Dave 2021-7-27

在 MATLAB Online 中打开

Not sure where you're headed with the result, but there are a lot of solutions here that would skip the for loop altogether:

%% Alternative 1
ind = data(:,2) == 5; % Logical index of second column == 5
hit = find(ind);    % Row number where second column == 5
L1 = nan(sum(ind),1); % Initialize L1 to have an item for each hit
L2 = nan(sum(ind),1); % Initialize L2 to have an item for each hit
L1(hit>1) = data(hit(hit>1)-1, 2); % L1, where the hit row is bigger than one, should get the previous row, second column
L2(hit>2) = data(hit(hit>2)-2, 2); % L2, where the hit row is bigger than two, should get two rows up, second column
% display the rows that don't have a 5:
miss = find(~ind);
disp(compose("Row %d has no number 5 (in the second column)", miss))
%% Alternative 2 
L1 = [nan; data(:,2)];          % Pad the second column of data with a NaN
L2 = [nan; nan; data(:,2)];     % Pad the second column of data with two NaNs
ind = data(:,2) == 5;           % true where data's second column is a 5
L1 = L1(ind);                   % Now just keep the items that were relevant
L2 = L2(ind);
% disp for the misses is the same as above
% * your code checked whether the second column had aa 5, if you wanted to
% check if *any* item in the row is a 5, try any(data==5, 2) for ind