cell strings to matrix

So i had to take a break from this code. I am completely new to cells and strings in matlab. I am trying to get this text file to be broken down to something that i could manipulate. Right now i have been able to isolate the data that i want to use. My problem is two fold:

1. The data file that i use is based on experimental data. Sometimes, the subject doesn't respond to the stimulus and the experimental recorder counts puts that in as MISSED RESPONSE or RESPONDED TO NOTHING or something of that nature. I have isolated the data and the MISSED RESPONSE or RESPONDED TO NOTHING get stored in separate cells in the cell array. Is there any way where i could make them one or have matlab look at any NAN value and discard that whole data row.

2. Can i convert it to, say, a matrix. This code is going to be a part of a bigger data analysis code and most of the data analysis part in it is through matrix. I am guessing that if i keep the data from this code in matrices, it might prevent any inconsistencies that might pop up later on. Although, i might be completely wrong. The code is below, so far, and i am also attaching the text file that i am working on right now.

[filename] = uigetfile('*', 'Select the data file')

import_data = importdata(filename)

[data] = import_data.data

fid = fopen(filename,'r')

scan_init_cell = textscan(fid, '%s %s %s', 'headerlines',20, 'CollectOutput', 1,... 'TreatAsEmpty',{'Responsed','to','Nothing','Missed','Response'})

fclose(fid)

scan_init = [scan_init_cell{:}]

scan_init = scan_init(1:find(~cellfun('isempty',strfind(scan_init,'''''''''''''''''''''''''''''')),1)-1,:)

scan_init_cell = cellfun(@isnan, scan_init, 'un', 0);

Answer 1

[Star Strider]

I remember this from one of your previous questions, but not all the details. To convert the entirely numeric data in ‘scan_init_cell’ to a matrix, use the cell2mat function.

If it is not entirely numeric, you likely have to keep it as a cell.

[Sanwal Yousaf]

yes, i took this on at this research facility as a challenge because no one else seemed to work on it, i am starting to see why. I took a week off from this code just to refresh myself. Is there any way i could just discard the entire row of the cell that has a NAN, but becuase of each word in the NAN being stored in a different cell, it is making it hard for me to do it

[Star Strider]

I must be missing something. When I went back and found the code I posted to your Question on 2016-01-08 (that I fortunately saved):

% % % % % URL: http://www.mathworks.com/matlabcentral/answers/263164-getting-only-numbers-from-an-88x3-cell-array
fidi = fopen('Sanwal Yousaf USE_THIS.txt','r');
scan_init_cell = textscan(fidi, '%f%f%f', 'HeaderLines',20, 'CollectOutput',1, 'TreatAsEmpty',{'Responsed to Nothing','Missed Response'});
fclose(fidi);
scan_init = cell2mat(scan_init_cell);
scan_init = scan_init(~isnan(scan_init(:,3)),:);            % Omit Rows With ‘NaN’ In Column #3

I found that the ‘scan_init’ variable is a (38x3) double array (entirely numeric matrix) with all the NaN rows removed. It worked for me with the ‘USE_THIS.txt’ file you uploaded. If you are having problems using my code with other files, I would need one of those files to see what the problem is.

[Sanwal Yousaf]

you, sir. I LOVE YOU. NO exaggeration. I can't say when i never used this. It works! Praise the maker

[Star Strider]

As always, my pleasure!