Classification Rule Builder behaviour

[sabined51282667]

Hi community, 

we are experiencing an unexpected behaviour with the classification rule builder. The rule consists of 107 different expressions and it is based on a always 9 digits long code.

Important to say: This rule hasn't been changed since months. And also the classification names didn't change. 

Today we noticed some unclassified values and found that, when testing the rule set, it is giving us this result: 

After this testing, Adobe is telling us all the rules are not working: 

But if we activate it again, it is giving us a success result: 

In Workspace we see that all the Codes that have been classified once, are still fine. But Codes that are coming in new are not getting classified anymore. Did anyone find this behaviour before and knows what to do (other than contacting Client Care :D)

Best regards

Sabine

[MandyGeorge]

You said you've been using this for a while. Has the format of your values changed?

Have you used an external regex tester to make sure the syntax works? I copied the regex from your screenshot and the test key, and put it into regex101.com. It said that it didn't find a match. I'm not an expert in regex, but \d should be a digit between 0-9. Since your string doesn't start with a number, it starts with a letter, that could be why it isn't matching. 

I would suggest doing some testing with your regex expressions and some test keys in a platform outside of Adobe to make sure that it's capturing what you want it to. 

[Jennifer_Dungan]

That is strange indeed... I haven't seen this before (you are making me worried about some of my rules now).

While I am slightly confused about what looks like the same Regex over and over again in the limited window, they look like valid regex to me....

Without knowing more about your setup (seeing all 107 rules, seeing data that these rules are processing, etc), I don't know if I can offer much in the way on concrete suggestions...

I know that contacting Client Care can be a pain, but they do have access to your suite and data, and they can review all the rules, as well as often have access to server logs that might give more information about the error (beyond the so called "user friendly" error that doesn't tell you much about the actual problem)... 

[bjoern__koth]

Maybe 107 is the number when things start to break

Seems like a very high number but you surely have your reasons.

If neither the rules nor the format has changed, maybe reach out to support.

Indeed an odd behavior

[fhusain]

I recently discovered that if I regex contains '\z', it creates an issue; when it is replaced with '\Z', you no longer get an invalid regex error. Not sure if this one is related, but you can if any of your rules contain '\z'. 

[Jennifer_Dungan]

Well that is odd.... \z should just be a literal "z" and \Z should be the same but for an uppercase Z.. but since "z" isn't a reserved shorthand, or a special character, there probably isn't any need to use "\z", you can just use "z"... but I still think that is very random.. good to know.

One more thing, Adobe Regex doesn't like "0" in isolation... those also fail... even though 0 should be a completely valid character....

[sabined51282667]

Thank you all for your answers. We recreated the whole rule now with the same entries and now it is working. No idea what made Adobe hickup here. We didn't have any \z or a isolated 0 in the rule.