Use regexp in Matlab to return the value of a variable from a text file

I am reading a text file into Matlab called 'test.txt' which is structured as follows:

$variable1 = answer1; 
$variable2 = answer2;
$variable3 = answer3;

I read the text file into Matlab line by line using the following segment of code:

fid = fopen('test.txt.');
tline = fgetl(fid);
tracks = {};
while ischar(tline)
    tracks{end+1} = regexp(tline, '(?<=^.*\=\s*)(.*)(?=\s*;$)', 'match', 'once');
    tline = fgetl(fid);
end
fclose(fid);

This piece of code returns the value of each variable line by line and would output:

answer1
answer2
answer3

What I want to do is modify my regexp expression so that I can specify the name of the variable to retrieve and have Matlab output the value assigned to the variable specified.

E.g. If I specify in my code to find the value of $variable2, Matlab would return:

answer2

Regards

[Oleg Komarov]

Post few lines of your real text file and tell what you want to retrieve. Your regular expression doesn't make sense to me.

Answer 1

[Walter Roberson]

var2find = 'variable2';
tracks{end+1} = regexp(tline, ['(?<=^\$' var2find '\s*=\s*)([^ ;]*)(?=\s*;+\s*$)'], 'match', 'once');

Note: this assumes that the values do not contain embedded blanks or semi-colons, but does not assume that the value is present. For example,

$variable2 = ;

would be treated as an empty response.

I have also tweaked the syntax accepted so that multiple semi-colons are accepted in case of typos:

$variable2 = 345;;

However, the expression would have trouble with

$variable2 = 345; ;

If you want that to be acceptable you need to say whether the value should be '345;' or '345'.

I did not have difficulty reading the original regex, but I had to make guesses about what you wanted to have happen in case of extra semi-colons.

Answer 2

[Ken Atwell]

I cannot really follow your regular expression, but the following should help you get started:

 text = ['$variable1 = answer1;', ... 
    '$variable2 = answer2;', ...
    '$variable3 = answer3;'];

 varToFind = '\$variable2\W+=\W+(.*?);';
 found = regexp(text, [varToFind], 'tokens' );
 found{1}

Answer 3

[Andreas Justin]

Since it's unanswered, and no answer is accepted. this one using java.util.Properties(), which propably is a nice way to get variable by names.

so this is for future search results.

%%preparing
fid = fopen(fullfile('D:\','test1.txt'),'w');
txt = {'#Variables','a = 1','b = false','c = true','d = nope','e = 1e6'};
txt = regexprep(txt,'(.*)','$1\n');
txt = [txt{:}];

fprintf(fid,txt);
fclose(fid);
clear fid

%%getting startet
path = fullfile('D:\test1.txt');
fr = java.io.FileReader(path);
javaProp = java.util.Properties();
javaProp.load(fr);
fr.close();

%%get one specific
javaProp.getProperty('a')
javaProp.getProperty('c')
javaProp.getProperty('d')
javaProp.getProperty('e')

%%get all in arrays
c = javaProp.entrySet.toArray
for ii = 1:numel(c)
    keys{ii} = c(ii).getKey;
    vals{ii} = c(ii).getValue;
end

Answer 4

[Simon]

Hi!

Maybe there is no need for a regexp. Read in the file, separate all lines at "=" and find the desired variable in the left hand part:

% read in file
fid = fopen('test.txt');
FC = textscan(fid, '%s', 'delimiter', '\n');
fclose(fid);
FC = FC{1};

% find '=' to split lines
ind = (strfind(FC, '='));
% split in left and right hand part
lhp = strtrim(cellfun(@(x,y) x(1:y-1), FC, ind, 'un', 0));
rhp = strtrim(cellfun(@(x,y) x(y+1:end), FC, ind, 'un', 0));

% search for string in left hand part (handles even substrings)
tf = ~cellfun('isempty', strfind(lhp, 'able1'));
% show match
rhp(tf)