Matching a sequence of non-empty, delimited strings with a Regular Expression

Question

I'm not an expert with Regular Expressions, and I'm having serious problems matching a particular pattern.

The pattern is:

A sequence of consecutive, arbitrary words marked with a prefix and a suffix. Inside the word there should be at least one character.

I mean, suppose that the prefix is "AB" and the suffix is "YZ". With this input:

AB----YZAB====YZABYZ//AB++YZ,,,AB====YZAB---YZ

The matched groups should be:

AB----YZAB====YZ , AB++YZ , AB====YZAB---YZ

The group ABYZ should not be matched, because it is "empty" (there is nothing between the prefix and the suffix.

I tried with

(AB(.*?)YZ)+

But the ABYZ is detected as part of the sequence, as the "*" may match nothing. If I force to use non-empty groups with

(AB(.+?)YZ)+

But still no lock, it detects groups

AB----YZAB_____YZABYZ//AB++YZ and AB====YZAB---YZ

I tried many other, more complex, regExps, with no luck.

Any help would be very appreciated!

score 1 · Accepted Answer · answered Jan 24 '19 at 17:17

1

You may use

(?:AB(?:(?!AB).)+?YZ)+

Details

(?:AB(?:(?!AB).)+?YZ)+ - one or more repetitions of
- AB - an AB substring
- (?:(?!AB).)+? (or (?:(?!AB|YZ).)+) - any char but a line break char, 1 or more repetitions, as few as possible, that does not start an AB char sequence (a so-called tempered greedy token)
- YZ - a YZ substring.

answered Jan 24 '19 at 17:17

Wiktor Stribiżew

1

Works like a charm! Thank you very much for your help, and for the reference to the tempered greedy token explanation. – Yeray Rodriguez Jan 25 '19 at 08:03

1 Answers1