Problem using regexp to extract certain lines

I am trying to extract lines that begin with /keylog/midi from a file that looks like:

/keylog/midi 144 60 72 1.001300

/keylog/oscp 144 60 0.006736 1.030209

/keylog/oscp 144 60 0.000000 2.852801

/keylog/oscp 144 60 0.000000 2.869148

/keylog/midi 144 60 0 2.870843

And I need to separate the two lines from each other. The code I have right now is:

Fid=fopen('keyData.txt'); 

myLines=fgetl(Fid);

while ~feof(Fid)

      myLines=fgetl(Fid);

      on=regexp(myLines,'(/keylog/midi)(\s\d+\s\d+\s[^0]\s\d+(.)\d+)', 'match');   % on must equal the line that's fourth line is greater than 0 

      off=regexp(myLines, '(/keylog/midi)\s\d+\s\d+\s[0]+\s\d+(.)\d+', 'match'); % off must equal the line that's fourth number is 0

end

When I run the code

off='/keylog/midi 144 60 0 2.870843' 

on={}

What is wrong with my regexp for on?

Answer 1

[Simon]

Hi!

I would suggest another approach. It is (in my opinion) easier to debug, because you can track your steps easily.

% open and read file
fid = fopen(FileName);
FC = textscan(fid, '%s', 'delimiter', '\n', 'whitespace', '');
fclose(fid);
FC = FC{1};

% remove blanks on start and end of line
FC = strtrim(FC);

% find all lines with '/keylog/midi'
FCmidi = FC(strncmp('/keylog/midi', FC, 12));

% read each remaining line, skipping the string '/keylog/midi'
C = cellfun(@(x) sscanf(x, '%*s %d %d %d %f'), FCmidi, 'UniformOutput', false);
% format C: each column is one log entry
C = [C{:}];
% on/of flag is in row 3
onoff = C(3, :);

Answer 2

[Walter Roberson]

[^0]\s should be [^0]\d* in order to eat the digits after the first non-zero one (e.g., [^0] will match the 7, and then the \d* will match the 2.

In the off expression, [0]+ will match one or more 0's. Will there ever be multiple 0's there, such as 00 ? If not then it would make more sense to get rid of the + and change the [0] to just 0

[John]

I tried it, and it didn't work. It still gives me { } when I run the code.

[Walter Roberson]

Note: use \. to indicate a literal period.

Could you show your modified regular expressions?

[John]

This is the code after I added the \d* after the [^0], and changed the (.) to \. The first one still does not work.

 on=regexp(myLines,'(/keylog/midi)\s\d+\s\d+\s[^0]\d*\s\d+\.\d+', 'match');
      off=regexp(myLines, '(/keylog/midi)\s\d+\s\d+\s0\s\d+\.\d+', 'match');

Answer 3

[Yamoussa SANOGO]

Hi there, I know this question has been around for a while, but I would add my suggestion in the case somebody else has the same problem. My approch would be a simple lookahead like this :

text =

" /keylog/midi 144 60 72 1.001300

/keylog/oscp 144 60 0.006736 1.030209

/keylog/oscp 144 60 0.000000 2.852801

/keylog/oscp 144 60 0.000000 2.869148

/keylog/midi 144 60 0 2.870843 " ;

rule = '(?<=\/keylog\/midi)(\s*\d*\s*\d*\s*\d*\.?\d*\s*\d*\.?\d*)' ;
matched_data = regexp(text,rule, 'match');

Then convert the matched data to string :

matched_data = [matched_data{:}];

This approch can be generalized by making the prefix '/oscp' and '/midi' a string variable and concatenate with the rest of the matching rule.