Hi All,
I just have some issues with the Tertiary Values and Classification Rules that we have setup and I'm wondering if you have encountered the same.
Maybe you guys may also have insights on what have we been setting wrong on our Classification Rules
The setup is like as below:
eVar X = <item 1>|<item 2>|<item 3>
Sample Values can be:
The Classification Rules as setup as
So when we created a report for Rule 3 with a custom event, the expected value should be just "yes" or "no".
The issue is, our report shows "yes", "no" and "unspecified".
When I breakdown the "unspecified" by eVar X, it will show the correct values with only "yes" or "no" values. No empty one.
Solved! Go to Solution.
Views
Replies
Total Likes
hmm ok.. is it possible that there are spaces (or worse.. a non-space character that you can't see)? Or maybe a case issue... I know that Adobe Analytics tracks the data with cases... but a while ago, they did start to group the upper and lower case variants under the version first seen... but I wonder if the underlying data is actually a different case (even though you can't see it)
I assume your rules are using regex?
When I did a sample test... and I made one or more of the values in the sample Key uppercase, this is what I see:
The "Match Result" isn't highlighted in white indicating a match, and none of the "Match Groups" are there....
When I change the sample key to all lowercase, I get this:
Unfortunately, Adobe's regex rules are rather simplified, and you cannot add the "ignore case" flag.....
So I wonder if a more generic rule like (\w*)\|(\w*)\|(\w*) would work better? (though this doesn't allow for potential white space in the middle of the values if that happens to be the issue)
Notice I put the second "Yes" back to a capital, but the rule still pulls out the value.....
I am only guessing at your regex at this time.. maybe you already accounted for this.
Views
Replies
Total Likes
Additional Notes:
We have tested also the rules on the Classification Rules Builder and all are correct.
Views
Replies
Total Likes
Classification can take some time to be classified.
Could it be you are running a report for a period where keys did not have time to be classified ?
Agreed, processing rules only run every 4-6 hours... if you are looking at "today's data" everything that has been collected since the last classification process will be unspecified until the next job runs. Likewise, new rules only have a lookback max of 6 months... if you created rules today, then try to pull all of last year the new rules won't reach far enough back.
Hi @Alexis_Cazes_ and @Jennifer_Dungan,
Thanks for your quick responses.
However, the issue started last Jan. 10th 2023. If it's just with the processing time, ideally the unspecified ones posted last Jan. 10th should be classified now but I'm still seeing them in my report.
The last time I have updated my classification rules are 2 months ago. I'm still seeing this issue today. Will update at the end of the day if there will be any progress.
I would need to update my classification rules in the next week, would it be advisable to delay the classification rules updates until this issue has been clarified?
Thanks again for your advises.
Views
Replies
Total Likes
hmm ok.. is it possible that there are spaces (or worse.. a non-space character that you can't see)? Or maybe a case issue... I know that Adobe Analytics tracks the data with cases... but a while ago, they did start to group the upper and lower case variants under the version first seen... but I wonder if the underlying data is actually a different case (even though you can't see it)
I assume your rules are using regex?
When I did a sample test... and I made one or more of the values in the sample Key uppercase, this is what I see:
The "Match Result" isn't highlighted in white indicating a match, and none of the "Match Groups" are there....
When I change the sample key to all lowercase, I get this:
Unfortunately, Adobe's regex rules are rather simplified, and you cannot add the "ignore case" flag.....
So I wonder if a more generic rule like (\w*)\|(\w*)\|(\w*) would work better? (though this doesn't allow for potential white space in the middle of the values if that happens to be the issue)
Notice I put the second "Yes" back to a capital, but the rule still pulls out the value.....
I am only guessing at your regex at this time.. maybe you already accounted for this.
Views
Replies
Total Likes
I'd say that spaces are the primary suspect in this case.
You can also fix this with Classifications Import. And actually, I'd say that this is the easiest solution of all. Given your 3 items, you have a limited set of combinations of "yes" and "no". So create a classifications file that has all of those combinations, and set your classified values accordingly. Then import that file into Classifications Importer.
If you don't want to do that, you can also export whatever classifications that Classifications has already classified. (Sidenote: whoa, that was a mouthful!) You can then inspect the exported file to see why your 3rd item isn't being classified as expected.