How to ignore punctuation in a user string while scanning for words (textscan())?

It works perfectly now. thanks to oleg and lucas for ur help! if ur interested, here's how it looks like in the end:

function pushbutton1_Callback(hObject, eventdata, handles)
words = get(handles.editbox, 'string'); %scans user input string from editbox
wavdirectory = 'C:\Program Files\MATLAB\R2010b\Recordings\';
wordsstring = regexp(words, '\w+', 'match') ; %reads string only, ignores punctuation
[j, k] = size(wordsstring); %stores number of words in user input string
for m = 1:k
  thisfid = [wavdirectory wordsstring{m} '.wav'];
try
  [y, fs] = wavread(thisfid);
  sound(y, fs);
catch
  fprintf(1,'Failed to process file wave "%s" because: ', thisfid);
  lasterror
end
end

[Oleg Komarov]

How do you get the user-input? http://www.mathworks.com/matlabcentral/answers/6200-tutorial-how-to-ask-a-question-on-answers-and-get-a-fast-answer

Answer 1

[Lucas García]

This removes the specified punctuation in your word:

regexprep(word, '[-!,.?]', '')

[Jan Donyada]

does regexpprep return the result in a cell array like textscan?

[Lucas García]

Yes it does.

Try:

>> word = {'this.-is..!a,test', 'it!!.works???-'};

>> C = regexprep(word, '[-!,.?]', '');

>> class(C)

ans =

cell

[Jan Donyada]

ok i used textscan after using the results of regexp so that i didn't have to alter the for statements. now it works but i encounter a problem when a malay word like 'kawan-kawan' is keyed in. the '-' is not read, and the returned word becomes kawankawan. this results in the wav file not playing as the wav file is kawan.wav. any ideas?

[Oleg Komarov]

word = {'kwan-kwan'};

C = regexp(word, '[\w-]+', 'match')

[Jan Donyada]

hmm doesn't seem to work.. i've updated the code in the main question. it now successfully plays the .wav files even if the input string has punctuation, but in the case of 'kawan-kawan' it still does not read 2 separate 'kawan' but 1 'kawankawan' which causes wavread to not be able to read kawan.wav

[Oleg Komarov]

Ah ok I thought you needed kwan-kwan! Then Use the version in my post directly, it strips the '-' and creates 2 words.

[Jan Donyada]

mm yeah i did use the code in ur post (updated the code in the question to reflect it). it still only returns 'kawankawan' as one word. i tried a standalone test program to confirm that it did. what do you think?

[Oleg Komarov]

word = {'this.-is..!a,test', 'it!!.works???-kwan-kwan'};

C = regexp(word, '\w+', 'match')

C{2}

C =

{1x4 cell} {1x4 cell}

ans =

'it' 'works' 'kwan' 'kwan'

[Jan Donyada]

ah the array dimensions were slightly different. i modified the for loop so it works. thanks guys! updated code is above if ur interested

Answer 2

[Oleg Komarov]

Use regexp.

If you want more details provide some example inputs and required output.

From Lucas' example:

word = {'this.-is..!a,test', 'it!!.works???-'};
C = regexp(word, '\w+', 'match')

[Jan Donyada]

ok i included the code within the question.

[Lucas García]

Nice, Oleg!

Answer 3

[Walter Roberson]

Ah, the poster broke this out in to a separate question, which I did not see before I answered in the original thread. My answer there was:

Before the textscan:

words = lower(words);
words(~ismember(words, ['a':'z' ' '])) = ' ';

then go ahead with the textscan

On second look, that could be shortened to

words = lower(words);
words(~ismember(words, 'a':'z')) = ' ';