MATLAB Answers

Regexp: different behavior for the same type of expressions

Asked by Giorgos Papakonstantinou on 19 Sep 2013

I want to capture everything except the tokens:

name, '_' and '.iv2'

name =
A7122
 >> filename'
 ans = 
    'A7122_60a.iv2'
    'A7122_60b.iv2'
    'A7122_70a.iv2'
    'A7122_70b.iv2'
    'A7122_90a.iv2'
    'A7122_90b.iv2'
    'A7122_100.iv2'
    'A7122_120.iv2'

I do this:

str=regexp(filename, [ '(?:[^' name '_])\w*(?:[^.iv2])' ], 'match');

And the answer is the following!

>> celldisp(str)
str{1}{1} =
60a
str{2}{1} =
60b
str{3}{1} =
0a
str{4}{1} =
0b
str{5}{1} =
90a
str{6}{1} =
90b
str{7}{1} =
00
str{8} =
       {}

I don't understand why regexp has a different behavior for i.e. in filename(1) and filename(3)

  2 Comments

Cedric Wannaz
on 19 Sep 2013

Your mistake is that [^A7122_] doesn't stand for "any six letters expression that is not 'A7122 _'", but instead for "any character which is not in the pool of literals {'A', '7', '1', '2', '_' }". The same applies to [^.iv2]. This is why 70a and 70b for example are not matched, but you get instead 0a and 0b.

Thank you very much Cedric! Very nice explanation!

Tags

Products

No products are associated with this question.

2 Answers

Answer by Vishal Rane
on 19 Sep 2013
Edited by Vishal Rane
on 19 Sep 2013
 Accepted answer

You can use:

regexprep( filename, [ name, '_|.iv2'], '')

Also

regexp( filename, [ name, '_(\w*).iv2'], 'tokens')

  1 Comment

Thank you Vishal! Much more compact syntax! However, do you know why regexp has this behavior?


Answer by Andrei Bobrov
on 19 Sep 2013
Edited by Andrei Bobrov
on 19 Sep 2013
str = {'A7122_60a.iv2'
    'A7122_60b.iv2'
    'A7122_70a.iv2'
    'A7122_70b.iv2'
    'A7122_90a.iv2'
    'A7122_90b.iv2'
    'A7122_100.iv2'
    'A7122_120.iv2'}
cl = regexp(str,'(?<=_)\w*(?=\.)','match');
out = cat(1,cl{:});

  2 Comments

Thank you Andrei! Do you know why is this happening?

Andrei Bobrov
on 20 Sep 2013

Please read about regexp, parts:

- Regular expression :

about

Metacharacters ( \w ),

Quantifiers ( expr* ),

Lookaround Assertions ( expr(?=test) and (?<=test)expr )


Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply today