regexp string to numeric array

6 views (last 30 days)
Knut on 4 Nov 2016
Edited: Jan on 4 Nov 2016
I have what seems like a common problem: a series of strings containing numeric data mixed with text. I want to extract those numbers into vectors or arrays. I have come up with the mock-up below, but it feels like a cludge. Is there a simple, readable one-liner that does the same?
str = {'bob22alice666buster2', 'donald42lisa00buddy9'};
pat = '\w*(\d{2})\w*(\d{2,3})\w*(\d{1})';
for idx = 1:2
tmp = regexp(str{idx}, pat, 'tokens');
male(idx) = str2double(tmp{1}(1));
female(idx) = str2double(tmp{1}(2));
doggie(idx) = str2double(tmp{1}(3));
I was hoping for something ala:
for ...
Mx3arr = str2num(regexp(str{idx}, pat, 'tokens')');

Answers (1)

Jan on 4 Nov 2016
Edited: Jan on 4 Nov 2016
str = {'bob22alice666buster2', 'donald42lisa00buddy9'};
pat = '\w*(\d{2})\w*(\d{2,3})\w*(\d{1})';
tmp = regexp(str, pat, 'tokens');
C1 = cat(1, tmp{:});
C2 = cat(1, C1{:});
M = str2double(C2);
I still prefer the dull string parsing without regexp:
S = sprintf('%s ', str{:});
S(isstrprop(S, 'alpha')) = ' ';
M = sscanf(S, '%d', [3, inf]);
Although it looks less smart, it works without clutter and much faster.


Find more on Characters and Strings in Help Center and File Exchange


Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!