How to read data after a specific word from txt file?

I just start learning matlab. Now I want to read data from a txt file. The txt looks like this

 EvtTime 12:11:33.034.438   
 Raw     0.00  m65S2c6     0.85  S1    22.43  S2    22.09  S3   -72.46 
 TTime 41:00:00.000.000  m65S2c7     0.02  S1    22.78  S2    23.35  S3   -44.46 
 TEHi     0.00 m65S2c11    -0.01  S1    22.66  S2    23.18    S4T03    22.15 
 TotalTm   160.38    S1T01    20.05    S1T11    23.55    S5T12   -11.18   S6BV08    42.26 
  LiveTm   114.63    S1T02    20.73    S1T10    20.33    S5T10   -15.52   S6BV05    42.13
 EvtTime 12:15:23.225.518   65S2c7     0.01  S1-5C24   111.06  S2    23.41  S3    24.21  
 TotalTm   173.38    S1T01    21.95    S2T10    22.67    S3T19   -60.22   S4BV04    35.83 
  LiveTm   154.32    S1T02    21.70    S2T11    21.68    S3T20   -54.67   S4BV05    35.92

I need the value of 'EvtTime', 'TotalTm' and 'LiveTm', and the value is the number next to the words. The values should be read in vector. It will be like TotalTm = [160.38 170.38], and the LiveTm = [114.63 154.32]. The EvtTime will be read as time to plot. My code doesn't work well. Here is my code:

fid = fopen('data.txt','r');
      while ~feof(fid)
      st = fgetl(fid);
       if  ~isempty(findstr(st,'TotalTm'))
          ftell(fid);
          fseek(fid,-9,0);
          F = cell2mat(textscan(fid,'%5d')) 
       end
       st = fgetl(fid);
       if  ~isempty(findstr(st,'LiveTM'))
          ftell(fid)
          g = cell2mat(textscan(fid,'%5d'))
       end
      end

The value it read is not correct.And I don't know how two make the 'EvtTime' as a vector. Could you please help me to solve the problem ? Many thanks in advance.

Have a gread day

Ashley

Answer 1

[Joel Miller]

Hi, I think the code below will do what you want. It uses textscan on the line you have read instead of the entire text file; it creates column vectors for TotalTm and LiveTM and a column vector of strings for EvtTime. Using 'elseif' allows you to check each line for the three strings you are searching for.

F = [];
g = [];
h = [];

fid = fopen('testtext.txt','r');
      while ~feof(fid)
      st = fgetl(fid);           
       if  ~isempty(strfind(st,'TotalTm'))
         stread = textscan(st,'%s %f');
          F = [F; stread{2}(1)];
       elseif  ~isempty(strfind(st,'LiveTm'))
         stread = textscan(st,'%s %f');
          g = [g; stread{2}(1)];
       elseif ~isempty(strfind(st,'EvtTime'))
         stread = textscan(st,'%s %s');
         h = [h; stread{2}{1}];
       end
      end
fclose(fid);

You can convert the EvtTime strings into whatever format is convenient for plotting.

[AUWAL ABUBAKAR]

Thanks Joel. Your contribution really works for me. However, I am running the script over multiple files using a loop. i find it difficult to collect each output as a vector. eg. F should be 10 values if i run the script over 10 files. the same goes for g and h

Any help will be appreciated

Thanks

Answer 2

[Joel Miller]

Auwal, if I have a series of files in a directory, e.g. data1.txt, data2.txt, data3.txt, etc., I can use the directory command to get a structure with the file names, then loop the code above to put the results from all the files into vectors F, g, and h. Is this what you are looking to do?

F = [];
g = [];
h = [];
datafiles = dir('data*.txt');

for k = 1:length(datafiles)
    fid = fopen(datafiles(k).name,'r');
          while ~feof(fid)
          st = fgetl(fid);           
           if  ~isempty(strfind(st,'TotalTm'))
             stread = textscan(st,'%s %f');
              F = [F; stread{2}(1)];
           elseif  ~isempty(strfind(st,'LiveTm'))
             stread = textscan(st,'%s %f');
              g = [g; stread{2}(1)];
           elseif ~isempty(strfind(st,'EvtTime'))
             stread = textscan(st,'%s %s');
             h = [h; stread{2}{1}];
           end
          end
    fclose(fid);
end

[AUWAL ABUBAKAR]

Thanks Joel, it works fine.

One more thing please. What if I want to extract three values thant are next to the word, all the values are seperated by commas.

Using the this solution I only get the first value.

thnks

[Joel Miller]

Auwal, Akira's solution is preferable if your version of Matlab supports it. If you have an older version of Matlab, you can read additional numbers with a single call to textscan:

stread = textscan(st,'%s %f %f %f');
F = [F; stread{2}(1), stread{3}(1), stread{4}(1)];

Answer 3

[Akira Agata]

If your MATLAB is R2019a or later, you can use readcell function.

Also, by using indexing technique, you can do your task faster, like:

fileList = dir('data*.txt');

EvtTime = [];
TotalTm = [];
LiveTm = [];

for kk = 1:numel(fileList)
  C = readcell(fileList(kk).name,'Delimiter','\r\n');

  idx = startsWith(C,'EvtTime');
  EvtTime = [EvtTime; regexp(C(idx),'[\d|:|.]+','match','once')];

  idx = startsWith(C,'TotalTm');
  TotalTm = [TotalTm; regexp(C(idx),'[\d|.]+','match','once')];

  idx = startsWith(C,'LiveTm');
  LiveTm = [LiveTm; regexp(C(idx),'[\d|.]+','match','once')];
end

TotalTm = str2double(TotalTm);
LiveTm = str2double(LiveTm);

[AUWAL ABUBAKAR]

Thanks for your solution, Akira

Regards