pull out number after specific string in txt file

Im struggling to find a suitable way to pull a specific value from a txt file.

I need the number after the wording "image quality score" , so in the text below, it would be the number 33.44468. Theres are many 1000 lines of text and about 100 lines containing this image quality score number.

14-12-15 14:33:18.962 +00 000000060 INF: Camera 1 Reg1 1 (E41G51_1) tile 45603 image quality score (red image): 33.44468

I have used the following, but not sure how to then extract an array of numbers that I then want to plot.

wordString='image quality score';
A=fileread(file);
B=strfind(A, wordString)

thanks for any help Jason

[Adam]

Using regexp may be your best method since you have extra stuff, i.e. '(red image): ' between the word you want to locate and the number you wish to pull out.

My wrestling with regexp took me quite a while to parse a line of a file according to some specification though so I'm afraid I don't have time to volunteer working out the exact expression!

I assume it isn't always the case that the number you want is always at the end of a line that contains the string "image quality score"? If it is then you can get away with just scanning all the lines, finding those containing your string and then just extract the last floating-point number on the line

[Jason]

I've got this approach nearly to work, not sure its the most elegant way:

fid = fopen(file,'r');
while 1
    tline = fgetl(fid);
    if strfind(tline, 'image quality score')>0
        U=strfind(tline, 'image quality score')
        tline(U:end)
        sprintf('%s   %f',tline)
    end    
    if ~ischar(tline)
       break
    end
end

[Jason]

Thanks Adam. ow do I extract last floating number??

Answer 1

[Guillaume]

As Adam said, a regular expression is probably the easiest (as long as you understand the regular expression language that is). There are many ways to construct the regular expression, depending on the allowed formatting for the number, etc. This will probably work for you:

filetext = fileread('somefile.txt');
numbers = str2double(regexp(filetext, '(?<=image quality score[^0-9]*)[0-9]*\.?[0-9]+', 'match'))

This will extract the first number after each of the 'image quality score' string. A number being defined as one or more digit (0 to 9) with an optional dot somewhere in there (but not at the end).

Complete explanation of the regex:

• look ahead for (i.e. start matching after) the string 'image quality score' followed by any number of characters not including '0' to '9', the (?<=image quality score[^0-9]*)
• once the preceding is found, match 0 or more characters from '0' to '9', the [0-9]*
• followed by an optional ., the \.?
• followed by one or '0' to '9' character, the [0-9]+

[Jason]

Adam and Guillaume, both of your answers are perfect for me. Can I accept both your answers?

[Adam]

Accept this one as it is more complete. Mine was more of a quick and dirty hack.

[Guillaume]

You can only accept one answer, but you can vote for the answer by clicking on the grey triangle

[Bret Kenny]

Hi Guillaume, I wondering if it is possible to match for negative numbers as well in this case? I am searching for xyz location coordinates in a text file and require the sign of the number. Appreciate any help you can offer.

Answer 2

[Adam]

If all lines containing image score have the same format you could use textscan on the tline that you extract (or infact just tline(U:end) as you have it) with the correct format specifiers to extract the tokens and then take the last one extracted.

Maybe easier, again if the format is consistent is:

tokens = strsplit( tline(U:end), ':' );
score = str2double( tokens{end} );

It isn't an elegant general solution, but you don't need a general solution if all cases have the same format.