Grouping in costume dimensions

[Ann9]

Hi,

I created a costume dimension that counts days since activation to current date and called it days_diff. 

Based on that dimension, I created another costume dimension called "age" where I want to group the number of days from activation so each group will include the ones that came before. 

The syntax I wrote: 

case(when(${day_diff}<7,"1 Week or less"),when(${day_diff}<30,"1 Month or less"),when
(${day_diff}<90,"3 Months or less"),when(${day_diff}<180,"6 Months or less"),when(${day_diff}<365,"1 Year or less"),"More than 1 year")

I ideally would like that "1 Year or less" will include all accounts in "6 months or less", "4 months or less" etc... however what actually happens is that "1 year or less" include only accounts that were activated 365 to 180 days ago. 

Can you please advise what can I do?

Thanks 

[Lobo-Google]

Hi @Ann9,

Have you tried using a custom bin or a custom group field? I wonder if that might be a good solution for your use case? There are some examples of using a custom bin and a using a custom group in the documentation that might also be helpful in determining if that might be a good thing to try!

[GavinW]

You are creating specific buckets there with your case statements, so each row can only fall in one of them. It sounds like you actually need to create filtered measures - you can create these in LookML, and you can also create them in the front end. If you are doing it in the front end, you can put your date_diff criteria in the 'Custom Filter' box in the bottom of the modal.