How do I filter time data from excel columns, where the values are numerical and the columns are character named?

Hello,

I'm trying to create a code that will filter eyelink fixations (labeled IA_FIRST_SACCADE_START_TIME) that took place before the stimulus onset (time). I've tried creating a variable that will only use the columns I need (6 and 8) like so

fixation = raw(6:8)

but it only outputs an array of 3 12s like so

[{12} {12} {12}]

I also tried creating a for loop as

for (i=='time'> 0) in raw %there was a stimulus present

if (i== 'time'> 0)

for ('IA_FIRST_SACCADE_START_TIME'> 'time')

add('IA_FIRST_SACCADE_START_TIME', fixations)

end

end

but I get the error for invalid use of operator

Finally, I've tried creating an indexed variable like

data=raw('time'>=0 && ('FIRST_SACCADE_START_TIME'>='time'==||));

but that comes out with another invalid use of operator error too

What should I do to make this work? Any help at all would be appreciated

Answer 1

[Simon Chan]

Better to look at the usage of those data import function used in MATLAB.

I use readcell in this case.

rawdata=readcell('RES_IAS (1).xlsx');                   % Use readcell to read all data
data=rawdata(:,6:8);                                    % Extract 6th to 8th column
header = rawdata(1,6:8);                                % Extract the headers
Start_time = cell2mat(data(2:end,1));                   % Convert to a matrix for data in column 6
time = cell2mat(data(2:end,3));                         % Convert to a matrix for data in column 8
idx1=Start_time>time;                                   % Find index for Start_time > time
result = [Start_time(idx1) time(idx1)];                 % This is the result you may want in matrix form
plot(Start_time(idx1))                                  % Plot the Start_time
hold on
plot(time(idx1))                                        % Plot the time
legend(header{1},header{3})                             % Legend

[Ruth Ronalter]

Thanks! Sorry it took a while to verify, my version doesn't have readcell as a function, had to use readtable, and then substituted table2array for cell2mat. It came out looking something like this

filename='RES_IAS (1).xlsx';                %Renamed file as filename
rawdata=readtable('filename')               %readcell undefined, used readtable instead 
data=rawdata(:,6:8);                        %Extract 6th to 8th column
header = rawdata(1,6:8);                    %Extract the headers
Start_time = table2array(data(2:end,1));    %since table and not cell array, used table2array instead to convert to array for data in column 6
time = table2array(data(2:end,3));          %converted to array for data in column 8
idx1=Start_time>time;                       % Find index for Start_time > time
result = [Start_time(idx1) time(idx1)];     % This is the result you may want in matrix form

There was an error outputting the legend on my version as well, subscripting a table using linear or multidimensional indexing wasn't supported. Reverted to using character legends for data.

figure; 
plot(Start_time(idxl))                      
hold on;
plot(time(idxl))     %added labels to axes and figure for clarity
xlabel('time stimulus presented (s)'), ylabel('time saccade seen(s)'), title('Filtered trials')  
legend('Saccade start', 'time stimulus presented') %added character titles for legend

[Simon Chan]

The function readcell on my previous code would read the header on the first row, and data starts from the second row.

If you use readtable, data will appear in the 1st row instead. On the other hand, the headers are stored in its Properties, hence you cannot just retrieve the first row as the header.

The following is for your reference if you would like to use function readtable:

rawdata = readtable('RES_IAS (1).xlsx');
Start_time = rawdata.IA_FIRST_SACCADE_START_TIME;
time = rawdata.time;
idx1 = Start_time>time;
result = [Start_time(idx1) time(idx1)];
plot(Start_time(idx1))
hold on
plot(time(idx1))
legend(rawdata.Properties.VariableDescriptions{6},rawdata.Properties.VariableDescriptions{8});

[Peter Perkins]

readcell has its purposes, but the right thing to use is MUCH more likely to be readtable.