Counting the number of a specific word in text.

I want to create a program, tah will count specific words in text. I wrote this:

filename='File name : ';
word= 'Searched word : ';
A=input(filename,'s');
B=input(word,'s');
count=0;
fid=fopen(filename);
C=textscan(fid,'%/s');
for i=1:lenght(C{1})
    words=char(C{1}(i));
    words=strrep(words,'.','');
    words=strrep(words,',','');
    if (strcmp(words,word))
        count=count+1;
    end
end

But this program isn't working.

I got:

Error using textscan
Invalid file identifier. Use fopen to generate a valid file identifier.

Error in count_word (line 7)
C=textscan(fid,'%/s'

[Adam Danz]

What's the value of FID? The error is telling you the problem: the file is not accessed properly. That could be because

• the file doesn't exist
• the file does exist but is not on the matlab path
• the file is protected
• the file is corrupt

[pawlo392]

Ok, I didn't see this:

fid=fopen(filename);

Should be:

fid=fopen(A);

But, Now I got:

Error using textscan
Unable to parse the format character vector at position 1 ==> %/s
Unsupported format specifier '%/'. See the documentation for TEXTSCAN for
supported formats.

[Guillaume]

Well, yes as the error message says, %/s is not a valid format specifier. The / has no place there.

Answer 1

[Jan]

Use uigetfile instead of an fragile input command. But even then you should check the output of fopen:

[fid, msg] = fopen(filename);
assert(fid ~= -1, 'Cannot open file %s: %s', filename, msg);

Instead of using the cells created by textscan, the import as char vector is easier:

[filename, pathname] = uigetfile;
file  = fullfile(filepath, filename);
word  = input('Searched word : ', 's');
str   = fileread(file);
str(ismember(str, '.,:!"''?')) = ' ';  % Replace punctuation
str   = [' ', str, ' '];
match = strfind(str, [' ', word, ' ']);
count = numel(match);

Answer 2

[Image Analyst]

Instead of

A=input(filename,'s');

try this:

[folder, baseFileName] = uigetfile('*.*');
filename = fullfile(folder, baseFileName)