How do I use REGEXP to match multi-digit values?

I've been reviewing the MATLAB Programmer's Guide in hopes of finding a solution to a current problem: How do I use REGEXP to match multi-digit values?

From earlier this year, I have the following commands:

str = 'Part ID: 1 or Part ID: 2 or Part ID: 5 or Part ID: 10';
exp = 'Part ID:\s+([25])\W';
tokens = regexp(str, exp, 'tokens');
Part_data = reshape(str2double([tokens{:}]), 1, []).'

These commands work great when I'm looking to match up single digit values like 2 and 5.

Now, let's use the following updated string and associated commands;

str = 'Part ID: 112 or Part ID: 220 or Part ID: 252 or Part ID: 106';
exp = 'Part ID:\s+([220])\W';
tokens = regexp(str, exp, 'tokens');
Part_data = reshape(str2double([tokens{:}]), 1, []).'

This results in Part_data = NaN

I get the same result when trying to match the value of 112.

If possible, how do I use REGEXP to match only these two multi-digit values?

[Cedric]

But what is the purpose? It seems to me that even if you were doing it right, you would essentially get 220, which is not a parameter or data (as you already know it). If you just want to see if there is a part 220, STRFIND might be more efficient. If you want to extract data that follow the header 'Part ID: 220', you should explain a bit more about the structure of the data, because it might be possible to extract it with a unique call to REGEXP.

Answer 1

[Azzi Abdelmalek]

str = 'Part ID: 112 or Part ID: 220 or Part ID: 252 or Part ID: 106';
exp = '(?<=Part ID:) 220';
m = regexp(str, exp, 'match')

Answer 2

[Kelly Kearney]

In a regular expression, [220] means "match a single 2, 2, or 0". Eliminate the brackets if you want to match a specific sequence:

 str = 'Part ID: 112 or Part ID: 220 or Part ID: 252 or Part ID: 106';
 exp = 'Part ID:\s+(220)\W';
 tokens = regexp(str, exp, 'tokens');
 Part_data = reshape(str2double([tokens{:}]), 1, []).'

If you want to match multiple sequences, use something like this:

 exp = 'Part ID:\s+(220|252)\W';

[Brad]

Right on. Thanks Kelly.

[Brad]

Kelly, one comment about your answer.

I successfully tried your last approach: exp = 'Part ID:\s+(220|252)\W';

In doing so, an idea came to mind. What if I wanted to match all of them except one? For example, if I wanted to match all but 106, I could go with the following:

str = 'Part ID: 112 or Part ID: 220 or Part ID: 252 or Part ID: 106';
  exp = 'Part ID:\s+(112|220|252)\W';
  tokens = regexp(str, exp, 'tokens');
  Part_data = reshape(str2double([tokens{:}]), 1, []).'

Resulting in:

Part_data =

   112
   220
   252

Do you know of a way to achieve the same result, using REGEXP, where I could use the equivalent of a logical NOT 106?

[Kelly Kearney]

There might be a way to do this with a singular regular expression, but I'd just capture them all and then filter out the ones you don't want:

 str = 'Part ID: 112 or Part ID: 220 or Part ID: 252 or Part ID: 106';
 exp = 'Part ID:\s+(\d+)';
 tokens = regexp(str, exp, 'tokens');

 Part_data = str2double([tokens{:}]);
 Part_data = Part_data(Part_data ~= 106);

[Brad]

Sounds good. Thanks.