Regular Expression, greedy over |

Question

for some weeks im working with regular expressions in php. Now my question: Is there any way, to make the RegEx greedy over | ?

for example subject: 012345abcdefghijklm

pattern: /(abcde|abcdefghi)/ will extract abcde, although abcdefghi is the greedier match.

The only way i found, is to sort the RegEx by the highest length of possibly matches

Thanks

did you want to capture `abcde` and `abcdefghi` into two separate groups? — Avinash Raj, Aug 02 '14 at 15:07
The example from above is very simple. In fact its a bit harder. Its not like the second is exactly the same, like the first expression at the beginning. The first is a complete different expression, that accidentally matches the second too. — guest1, Aug 02 '14 at 15:07
I just wanted to know, if there is something like a mode that captures the whole RegEx instead of finishing after the first match. (no, global is not what i meant) — guest1, Aug 02 '14 at 15:11
you could try this http://regex101.com/r/uI2oL4/1 to capture both strings in a seperate groups. — Avinash Raj, Aug 02 '14 at 15:15
Define "greedy over |". For this specific case it's better to rearrange. If you meant skipping matches when one alternation is matched use atomic groups. — Unihedron, Aug 02 '14 at 15:19

score 2 · Accepted Answer · answered Aug 02 '14 at 15:15

There is no way other than reordering the elements or by adding an optional non-capture group.

Regular Expression engines are eager. Because of the way the Alternation meta-character works (|), the first possible match exits the alternation clause.

Either re-order the possible choices (/(abcdefghi|abcde)/) or use an optional non-capture group (/(abcde(?:fghi)?)/).

score 0 · Answer 2 · answered Aug 02 '14 at 15:03

0

Here you should use:

abcde(?:fghi)?

answered Aug 02 '14 at 15:03

Andie2302

score 0 · Answer 3 · edited Aug 02 '14 at 23:28

0

Not a perfect solution. But it provides alternative other than the sorting one:

(abcde(?!fghi)|abcdefghi)

edited Aug 02 '14 at 23:28

HamZa

answered Aug 02 '14 at 15:03

HMK

1

This is slightly worse in performance over `/(abcdefghi|abcde)/` – Unihedron Aug 02 '14 at 15:18

3 Answers3