MATLAB: How to filter set of strings with set of patterns

cellarrayfilteringMATLABregular expressionstrings

I have a list of strings:

names = {'Robert', 'Bob', 'Charles', 'James', 'Jamie', 'Marie', 'Jamesbob'}

That I want to filter with a set of regular expression patterns:

patterns = {'.*ob', 'jam.*'}

How can I a list of the names which match with at least one of the patterns? (Without getting duplicates and while keeping the original order)

So would expect an output like:

{'Robert', 'Bob', 'James', 'Jamie', 'Jamesbob'}

I can find matches with a single pattern like:

regex_matches = regexpi(names, pattern{1}, 'match')

Can have multiple patterns without using a loop?

And here `regex_matches` is a cellarray of cellarrays. How would I turn this into a single level list?

(This questions sounds failry general but I couldn't find a previous question. Apologies if there is one.)

Best Answer

Try this

names = {'Robert', 'Bob', 'Charles', 'James', 'Jamie', 'Marie', 'Jamesbob'};
patterns = {'.*ob', 'jam.*'};
regex_matches = cell(numel(patterns), numel(names));
for i=1:numel(patterns)
    regex_matches(i, :) = regexpi(names, patterns{i}, 'match');
end
idx = any(~cellfun(@isempty, regex_matches));
matched_names = names(idx);

Result

>> matched_names
matched_names =
  1×5 cell array
    {'Robert'}    {'Bob'}    {'James'}    {'Jamie'}    {'Jamesbob'}

Related Solutions

MATLAB: Regexp : get all the match for one expression

It's not possible to do what you want with just a regular expression. A character can only be part of one match, so once you've captured 'bigFamily_' you can't capture it again as part of bigFamily_middleFamily_ within the same regular expression.

What I would do is just capture each *family part and rebuild all the possible combinations:

var =  'bigFamily_middleFamily_littleFamily_childName';
subfamilies = regexp(var, '[a-zA-Z0-9]+_', 'match');
combinations = num2cell(logical(tril(ones(numel(subfamilies)))), 2);
families = cellfun(@(c) strcat(subfamilies{c}), combinations, 'UniformOutput', false)

Or if you don't want the trailing '_':

var =  'bigFamily_middleFamily_littleFamily_childName';
subfamilies = regexp(var, '[a-zA-Z0-9]+(?=_)', 'match');
combinations = num2cell(logical(tril(ones(numel(subfamilies)))), 2);
families = cellfun(@(c) strjoin(subfamilies(c), '_'), combinations, 'UniformOutput', false)

MATLAB: Get substring from text

S = {'Hi B', 'Hi A', 'Hey guys D can''t come', 'Too Bad C', 'I Feel bad for D'};
names = {'A', 'B', 'C', 'D'};
namepattern = ['\<', strjoin(names,'\\>|\\<'), '\>'];
matched_names = regexp(S, namepattern, 'match');

The result will be a cell array the same size as S, with each entry containing a cell array of character vectors. The contained cell array will be empty if there are no name matches; it will contain a scalar cell array that contains a character vector if there is one match (the case you are expecting), and the contained cell array could contain multiple character vectors if multiple names were matched.

In the particular case where you are sure you only have at most one match, or you only want the first match if there are multiple matches, then you can use

matched_names = regexp(S, namepattern, 'match', 'once');

In that case, instead of matched_names being a cell array containing cell arrays, it would be a cell array directly containing character vectors

This code does not require that the names are single letters, and does not require that they are upper case. It will, however, only recognize them if they are isolated -- for example 'Too Bad C' it will detect that the 'B' in 'Bad' is followed by something that is word-like and so will not treat it as a match for the name 'B'. The code will consider digits and underscore to be word-like, so for example 'B_' and 'B3' would not be considered to match 'B' but 'B.' or 'B?' would be considered to match.

Best Answer

Related Solutions

MATLAB: Regexp : get all the match for one expression

MATLAB: Get substring from text

Related Question