regexp: either 'once' or 'all'

Hi, I want to do a search only for the first n times. But it seems the 'options' in regexp can only do 'once' or 'all'. no intermediate choice, say '5', huh? thanks a lot!

[Azzi Abdelmalek]

Give an example

[Anthony]

for example, a sentence like A='I have a dream have a dream have a dream bla bla bla...'. this sentence might be pretty long, so i wanna control for how many times it will search to make running time not too much. regexp(A,'dream','once') cannot accomplish this presumably, i can strsplit this into words using a cell, and do the regexp only for the first 40 words. something like B=strsplit(A,' '); regexp(B(1:40),'dream'); But, here comes the second reason: more important, this sentence belongs to a 10000*1 cell, each of which includes such a sentence. and i prefer to working with the entire cell array, not using any for loop. thanks a lot

Answer 1

[Walter Roberson]

(pattern){5}

In some cases what you might be after more might be

N = 5;   %times to match
Rexpat = [ repmat( ['(' pattern '[).*?]', 1, N-1), ')' ];
regexp(string, Rexpat)

[Cedric]

There should be a 3rd param. 'once' in the call to REGEXP, to avoid getting multiple matches if there are multiple times N occurrences of the initial pattern.

@Anthony: I would profile both the Rexmat pattern and the original pattern, because reducing the number of matches by increasing the complexity of the pattern doesn't always lead to an increase in performance. It's quite the opposite in fact.

Answer 2

[Image Analyst]

Perhaps you want strfind():

A='I have a dream have a dream have a dream bla bla bla...'; 
indexes = strfind(A, 'dream') % Find all locations.
indexes = indexes(1:5); % Take first 5 only.

In the command window:

indexes =
    10    23    36

[Image Analyst]

There's an old saying on how to write and speak: "Never say "Blah, blah, blah" when "Blah" will do. ;-)

[Cedric]

+1 for the saying-driven solution ;-)