How to search text file for string, and return the line?

Hello,

I need to find where in a file a string occurs, and grab the 5th and 6th values on the same line into 2 variables. Here's a sample of the data:

    401  FxTB                             2591.1675       15.3213     2569.5085     2619.7012       50.1926    kN
                                                                        20.840 s       4.080 s
                                                                           522           103  
    402  FyTB                                0.8813       17.3074      -67.7260       64.5470      132.2730    kN
                                                                        15.280 s      16.560 s
                                                                           383           415

I don't know where the FxTB will occur, but I need to store the 2569.5085 and 2619.7012 in two variables. I was using this:

       fmt = ['%*f %*s' repmat('%*f',1,2) '%f %f'];    % skip snb#/snbName/2numbers, then read 2 values: sensor 401
       v = cell2mat(textscan(fid2,fmt,1,'headerlines', 553-1));

...but the line doesn't always occur on line number 553.

Maybe I need to load the whole file into memory...since I do search for about 6 variables per file.

Answer 1

[Walter Roberson]

loading the file into memory is the easiest way.

If you split the loaded text into lines, then strfind() or regexp() can be used to do the searching.

With strfind() you would

mask = ~cellfun(@isempty, strfind(TextAsCells, 'FxTB'));
the_one_line = TextAsCells(mask);

You can do much the same with regexp()

mask = ~cellfun(@isempty, regexp(TextAsCells, 'FxTB'));
the_one_line = TextAsCells(mask);

However, you can also use regexp() in a way to extract the data values more directly, similar to the below.

If you do not split the loaded text into lines, then regexp() can be used to do the searching.

   f56 = regexp(TextAsSingleLine, '(?<=FxTB\s+(\S+\s+){2})\S+\s+\S+','match');

which would return {'2569.5085 2619.7012'}

Or you could use

parts = regexp(TextAsSingleLine, '(?<=FxTB\s+(\S+\s+){2})(?<f5>\S+)\s+(?<f6>\S+)','names')

parts = 
struct with fields:

    f5: '2569.5085'
    f6: '2619.7012'

to get it pre-split.

If it is always the same field numbers then you could use, for example,

parts = regexp(TextAsSingleLine, '(?<varname>(FxTB|FyTB|P7qT)\s+(\S+\s+){2})(?<f5>\S+)\s+(?<f6>\S+)','names')

which should match for variable names FxTB, FyTB, or P7qT with the matched name returned in the varname field of the resulting struct.

[David Pesetsky]

I'm having trouble loading the generally formatted file into memory to get your TextAsCells. Is there something other than fread I should use?

[David Pesetsky]

Ugh. I Tried:

text = fileread(MyFile);
parts = regexp(text, '(?<=FxTB\s+(\S+\s+){2})(?<f5>\S+)\s+(?<f6>\S+)','names')

...and it did find both values, but it seemed to take forever compared to my current method :(

[David Pesetsky]

Ah..this way is spectacularly fast. Then I can use the line number with my old method:

c=textread('my.txt','%s','delimiter','\n');
find(~cellfun(@isempty,strfind(c,'FxTB')))

Please comment.

[Walter Roberson]

One way:

text = fileread(MyFile);
TextAsCells = regexp(text, '\n', 'split');

Another:

fid = fopen(MyFile, 'rt');
TextAsCells = textscan(fid, '%s', 'Delimiter', '\n');
fclose(fid)

Using regexp is usually fast; it can slow down if there are a lot of false matches that it has to back-track over. Performance can be improved by using line anchors

text = fileread(MyFile);
parts = regexp(text, '(?<=^s+\S+\s+FxTB\s+(\S+\s+){2})(?<f5>\S+)\s+(?<f6>\S+)','names', 'lineanchors')

as then it knows that it does not need to keep looking in the line for occurrences of FxTB

Answer 2

[Image Analyst]

A very simple way is to use fgetl() ans strfind():

fid = fopen('fgetl.m');
tline = fgetl(fid);
lineCounter = 1;
while ischar(tline)
    disp(tline)
    if strcmp(tline, 'FxTB')
        break;
    end
    % Read next line
    tline = fgetl(fid);
    lineCounter = lineCounter + 1;
end
fclose(fid);

lineCounter is the line number where it occurred. Then you can parse tline to pull out any numbers you want from tline using sscanf() or textscan() or whatever.

[Kyle Boyce]

Very helpful. Thank you!

[Hamidreza]

This not working for a keyword in a line. only finds the works for whole line. What if we looking for only a key word?

[Image Analyst]

@Hamidreza

Use contains instead of strcmp():

if contains(tline, 'FxTB', 'IgnoreCase', true)
    break;
end

[taimour sadiq]

if contains(tline, 'FxTB', 'IgnoreCase', true)

%% How to replace this 'tline' containing 'FxTB' in file with some string 'Hello'

break;

end

[Voss]

@taimour sadiq:

filename_in = "test.txt";
filename_out = "test_modified.txt";

lines = readlines(filename_in);
lines(contains(lines,'fxtb','IgnoreCase',true)) = "Hello";
writelines(lines,filename_out);

% check the input and output files:
type(filename_in);

It was the best of times,
it was the blurst of times,
it was the age of FxTB,
it was the age of foolishness,

type(filename_out);

It was the best of times,
it was the blurst of times,
Hello
it was the age of foolishness,