Hey @WLS-DM
Thank You for the response.
We are only interested in the specific granularity defined in our formula — for example: brand_name + year for Brands, merchant_name + year for Retailers, and dynamic price band + year for Price Band calculations. Nothing beyond these groupings is required for our use case.
We also tried the approach you mentioned (denserank + sumover), but unfortunately we’re still not able to achieve the expected results, as highlighted above(as in the comment from May 5, 9:01 PM).
Here is the detailed formula we’re using — everything works as expected except for the Placements + Price Band combination, where we currently have a NULL placeholder. We’re looking to replace that part with a working solution that matches the logic used for Brands and Retailers.
You can also understand the intended granularity directly from the formula structure:
brand_name + yearfor Brands,merchant_name + yearfor Retailers,- and
dynamic price band + yearfor Price Band calculations.
Let me know if you see a clean way to plug this into the formula without disrupting existing behavior.
Alternatively, would it be better to create a static version of the dynamic price band directly at the dataset level and use that in distinctCountOver to ensure consistent results?
(The challenge is that our current Dynamic Price Band field is parameter-driven, making it difficult to replicate exactly at the dataset level.)
Also, I came across a related issue that seems quite similar:
DistinctCountOver not taking calculated field as partition by field
Regards,
Nikhil.