Regex is matching in Javascript but not PCRE

Question

I'm trying to match all fractions or 'evs' and strings (string1, string2) the following string with regex. The strings may contain any number of white spaces ('String 1', 'The String 1', 'The String Number 1').

10/3 string1 evs string2 8/5 mon 19:45 string1 v string2 1/1 string1 v string2 1/1

The following regex works in Javascript but not in PHP. No errors are returned, just 0 results.

(\d{1,3}\/\d{1,3}|evs).*?(.+).*?(\d{1,3}\/\d{1,3}|evs).*?(.+).*?(\d{1,3}\/\d{1,3}|evs).*?(.+) v (.+).*?(\d{1,3}\/\d{1,3}|evs).*?(.+) v (.+).*?(\d{1,3}\/\d{1,3}|evs)

Here's the expected result, other than group 6 and 7 (ran using Javascript):

If I add a ? to the first (.+) so that it becomes (.+?), I get the desired result but with the first string not captured:

As soon as I remove the ? to capture the whole string, there are no results returned. Can somebody work out what's going on here?

You should avoid +? and *?. Can you explain what you would like to achieve, we may think of an improved pattern? — Łukasz Nojek
– Łukasz Nojek, Commented Mar 12, 2020 at 20:14
@ŁukaszNojek Sure. I'm aiming to capture the fractions and the strings (string1 and string2), ignoring everything else. The regex string needs to be able to be built up dynamically by the columns that are defined in a prior function (each column has the regex for the element in that column e.g. a fraction is (\d{1,3}\/\d{1,3}|evs)). — Alex
– Alex, Commented Mar 12, 2020 at 20:44
Does (\d{1,3}\/\d{1,3}|evs) not give you the expected matches? — The fourth bird
– The fourth bird, Commented Mar 12, 2020 at 20:45
Can you provide the correct expected result? I'm still not sure if you want to match evs, string1 etc., — Łukasz Nojek
– Łukasz Nojek, Commented Mar 12, 2020 at 20:52

Wiktor Stribiżew · Accepted Answer · 2020-03-13 14:23:09Z

1

In PCRE/PHP, you may use

$regex = '(\d{1,3}\/\d{1,3}|evs)\s+(\S+)\s+((?1))\s+(\S+)\s+((?1))\s+(.+?)\s+v\s+(\S+)\s+((?1))\s+(\S+)\s+v\s+(\S+)\s+((?1))';
if (preg_match_all($regex, $text, $matches)) {
    print_r($matches[0]);
}

See the regex demo

The point is that you can't over-use .*? / .+ in the middle of the pattern, that leads to catastrophic backtracking.

You need to use precise patterns to match whitespace, and non-whitespace fields, and only use .*? / .+? where the fields can contain any amount of whitespace and non-whitespace chars.

Details

(\d{1,3}\/\d{1,3}|evs) - Group 1 (its pattern can be later accessed using (?1) subroutine): one to three digits, / and then one to three digits, or evs
\s+(\S+)\s+ - 1+ whitespaces, Group 2 matching 1+ non-whitespace chars, 1+ whitespaces
((?1)) - Group 3 that matches the same way Group 1 pattern does
\s+(\S+)\s+((?1))\s+ - 1+ whitespaces, Group 4 matching 1+ non-whitespaces, 1+ whitespaces, Group 5 with the Group 1 pattern, 1+ whitespaces
(.+?) - Group 6: matching any 1 or more char chars other than line break chars as few as possible
\s+v\s+ - v enclosed with 1+ whitespaces
(\S+) - Group 7: 1+ non-whitespaces
\s+((?1))\s+ - 1+ whitespaces, Group 8 with Group 1 pattern, 1+ whitespaces
(\S+) - Group 9: 1+ non-whitespaces
\s+v\s+ - v enclosed with 1+ whitespaces
(\S+)\s+((?1)) - Group 10: 1+ non-whitespaces, then 1+ whitespaces and Group 11 with Group 1 pattern.

edited Mar 13, 2020 at 14:23

answered Mar 13, 2020 at 14:14

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Alex Over a year ago

Thanks. I've updated the question to add the constraint that the words can contain any number of whitespace.

Wiktor Stribiżew Over a year ago

@AlexGodbehere The answer is still the same. I do not know your log format. You must use precise patterns for each field.

Alex Over a year ago

Thanks for your help. I ended up using .+.

(\d{1,3}\/\d{1,3}|evs) (.+) (\d{1,3}\/\d{1,3}|evs) ((?2)) (\d{1,3}\/\d{1,3}|evs) ((?2)) v ((?2)) (\d{1,3}\/\d{1,3}|evs) ((?2)) v ((?2)) (\d{1,3}\/\d{1,3}|evs)

Collectives™ on Stack Overflow

Regex is matching in Javascript but not PCRE

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related