Exclude a specific string from a RegEx expression in Google Analytics

Question

I have the following basic RegEx expression in Google Analytics that consists of a list of city names.

I would like to add another expression to the query such that a result is not returned if the matching text also has the word "job" in it.

Does this help? https://stackoverflow.com/questions/1687620/regex-match-everything-but-specific-pattern — BlackPearl, Jan 16 '20 at 16:18
@WiktorStribiżew: not so easy, because the lookbehind is not allowed to contain regular expressions inside of it. — Pierre François, Jan 16 '20 at 17:19
@Pierre Lookbehinds are regexes and in .net they can contain any length matching patterns. — Wiktor Stribiżew, Jan 16 '20 at 19:22
@BlackPearl I believe that this is the same suggestion made by Pierre Francois but it doesn't seem to work in Google Analytics. — zgall1, Jan 17 '20 at 16:11

score 0 · Answer 1 · answered Jan 16 '20 at 17:16

0

[Edited answer] This works:

^((?!job).)*(toronto|vancouver|calgary|halifax|winnipeg|ottawa|edmonton|regina|mississauga)(?!.*job)

It matches the city names only if "job" is not present before or after the city name.

See https://regex101.com/r/tXfBGf/3

answered Jan 16 '20 at 17:16

Pierre François

5,850
1
17
38

When I enter this into Google Analytics, I receive the following error: "Invalid regular expression. Your data request includes an invalid regular expression.". Thoughts? – zgall1 Jan 16 '20 at 19:13
GA regex engine is RE2 and it doesn't support lookarounds. – Wiktor Stribiżew Jan 16 '20 at 19:23
@WiktorStribiżew Is there an alternate solution that might work in GA? – zgall1 Jan 17 '20 at 16:12
@zgall1 Probably you may do that using exclude filters. – Wiktor Stribiżew Jan 17 '20 at 20:38

Devon Parsons · Answer 2 · 2020-01-16T17:34:20.047

-1

Provided your flavor of regex includes capture groups, this is easily solved using The Best Regex Trick, prior exclusion. Essentially:

"string".match(/(?:a|b|c)job|(a|b|c)/).match_group(1), where:
?:: Do not save the prior exclusion in a capture group;
(a|b|c)job are the strings to avoid capturing;
|(a|b|c) are the strings to capture.

This works because if the text matches the undesirable text, ajob, then the capture group is never populated. It will only be populated if the desired text a is matched without ajob

edited Jan 16 '20 at 17:34

answered Jan 16 '20 at 17:27

Devon Parsons

1,234
14
23

I am not able to follow how this concept can be applied to my specific case. – zgall1 Jan 17 '20 at 16:11
If you can write a regex that matches the text you don't want (e.g text matching /job/) then you can use this technique to convert that to matching on text you DO want while excluding those forbidden contexts. – Devon Parsons Jan 17 '20 at 16:34
Perhaps if you can clarify what you mean by "has the word 'job' in it" I can write a more specific regex to your case; for example, are you encountering text like `winnijobpeg`? I doubt it, so I used the simpler example of city followed by job a la `winnipegjob` – Devon Parsons Jan 17 '20 at 16:36
The word job is always separated from other characters by either a slash (/job/winnipeg) or a dash (job-winnipeg). – zgall1 Jan 17 '20 at 20:08
then use`/job[-\/](?:winnipeg|toronto|...)|(winnipeg|toronto|...)/`, match group 1. Sample here: https://regex101.com/r/jBfTeE/1 Pay close attention to the match group panel on the right side. – Devon Parsons Jan 20 '20 at 18:09

Exclude a specific string from a RegEx expression in Google Analytics

2 Answers2