The regex you posted will not match the key you posted. You can easily verify this by testing it in two different places in the rule builder, or online elsewhere (e.g. www.regex101.com). So,l if your classification report is getting popped with a value, then the value is coming from somewhere else.
Your posted CRB does not accurately reflect what you actually have
You have another rule in your CRB matching it
You imported classifications for that key via browser or ftp location
The key was previously classified from something else (manually or automated)
Sometimes there is funkiness with re-classifying columns vs. previously classified data. Especially if you rename columns instead of delete them, or delete them when they are not the last column listed. Did you recently re-classify the report? Try deleting all classification columns from the report and add them again.
I am having a similar issue. I ran a classification rule builder test. The results were perfect. All columns populated with expected parsed attributes - But when I run a report I only see Unspecified. The delimiter I am using is an underscore.
This is the only rule that classifies utm_campaign.
I'm not trying to classify any part of the above string - it should not match. However, it is matching with ‘mombaby_foc_17m8_care_hp-pn|video_|programmatic|w2045-twincities_demo_awareness|video_v1_ birth_|mombaby_foc_17m8_care_hp-pn’ to $1.
$1 is the equivalent of '^([^_|]*_[^_|]*_[^_|]*_[^_|]*_[^_|]*)$', the first parentheses.