Abbreviate regexp: match a if present, else match b

dear community,

can someone abbreviate the follwing if else in a single regexp?

str1='this is } and this is }, and this is }';
str2='this is } and this is }  and this is }';

if regexp(str1,'},','once')
    where=regexp(str1,'},')
else
    where=regexp(str1,'}')
end

where = 23

if regexp(str2,'},','once')
    where=regexp(str2,'},')
else
    where=regexp(str2,'}')
end

where = 1×3

9 23 38

i tried to adapt the example of the documentation, but was not able to make it work

(?(cond)expr1|expr2)

If condition cond is true, then match expr1. Otherwise, match expr2.

'Mr(s?)\..*?(?(1)her|his) \w*' matches text that includes her when the text begins with Mrs, or that includes his when the text begins with Mr.

regexp(str1,'}(,?)(?(1)},|})')

ans =

     []

best regards

Jonas

Answer 1

[Stephen23]

It is possible to combine them, but it will be fiddly and not very efficient because it requires in some way checking the entire text for the existence of the specified substring when performing every single match. Here is one approach that combines a conditional operator with a dynamic command:

st1 = 'this is } and this is }, and this is }';
st2 = 'this is } and this is }  and this is }';
rgx = '\}(?(?@contains([$`,$&,$''],''},'')),)';
regexp(st1,rgx)

ans = 23

regexp(st2,rgx)

ans = 1×3

9 23 38

Broken down:

% '\}(?(?@contains([$`,$&,$''],''},'')),)';
%  \}                                         literal curly brace
%    (?(                              ),)     conditional match comma
%       ?@contains(           ,      )        dynamic check text content:
%                  [$`,$&,$'']                entire input text
%                              ''},''         literal curly-brace & comma

You could probably do something similar with lookarounds. But as noted, the condition command will be called again and again with each match... this is rather inefficient, because your condition does not change. I do not see an obvious way to avoid this. A more efficient approach (if you are able to relax your requirements) is probably to check that condition once before calling regexp:

rgx = '\}(?(?@tmp),)';
tmp = contains(st1,'},');
regexp(st1,rgx)

ans = 23

tmp = contains(st2,'},');
regexp(st2,rgx)

ans = 1×3

9 23 38

[Jonas]

it seems like i misunderstood the concept of condition in this case since the condition is checked in each case (like you said). thanks for your code!