Classification help request

[#1]

Hi there,

Owing to some custom Adobe analytics instrumentation, we are not able to use the out of the box "Browser width" and "Browser height" dimensions.

As a consequence, we have enabled a new dimensions that allow us to capture the "browser width" and "browser height".

Here is an example of the value that we currently populate : "{""width"":1920,""height"":937}"

I would like to do two things :

• Make a classification in order to extract individually each value (one for browser width and one other for "browser height"
• Once the extraction is over, I would like to group these values under some buckets (one for browser width and one other for browser height) :
  • 100 to 199
  • 200 to 299
  • ...
  • 1000 to 1099

FYI, I already created the "traffic classification" then tried to build the rule via the "Classification rule builder" section but I'm currently struggling with the regexp.

Thank you in advance for your help !

[Fox_mulder007]

Hi @Fox_mulder007 ,

Part 1:

^"{""width"":([^,]*),""height"":([^,]*)}"$

sending $1 into the variable for width and $2 into the variable for height.

Part 2:

For this be sure to use sub classifications so that you can classify the already classified variable:

For 100-199,

^1[0-9][0-9]$

For 200-299,

^2[0-9][0-9]$

etc. mapping $0 to the corresponding range. 

If you don't want to use sub classification, you could just nest the two regular expressions and create the range from the initial value.

[Fox_mulder007]

Hi @jkm-disco,

I've just tried to replicate what you recommended and I'm not seeing the expected results : 

Issue #1 :

• Two rules have been created as you can see below : 

• When I used the "Rule set tester", it returned a good result : 

Issue #2 : 

• When I activated the rule then went to the dimensions that has been classified, I only saw "unspecified".

Did I do something wrong here?

Another question regarding your response for the part two please : 

We have ~500K unique values for this dimension, creating one regexp per grouping (eg : 0 to 99 , 100 to 199, ... , 10000 to 10099...) would be very time consuming. 

Is there any other solutions that would allow us to have this kind of grouping without heavy work ?

Thank you for your help,