So, keeping it simple, you want to know if the February being in the last 12 months is a leap year or not? Because your calculation can only return 2 numbers: 365 most of the time, and 366 when it's a leap year.

I believe the first thing is to decide if you want to calculate based on your incremental key or you want to calculate based on dates. The incremental key is something you (or somebody else) manage, and while now it's a date and would give the same result, functionally it's a different thing if you use that vs. doing a date calculation.

If you want to work on dates (the formula will always work, whatever your incremental key is), you can build that login by using native LSQL formulas.

Here an example of what it could be like (based on a column giving me a date, so I can make sure it does work correctly):

The best way to get there is by cutting your need into small pieces: get the first day of month of 1 year ago, then get the last day of the previous month, then calculate the difference in days between the 2 dates.

TIMESTAMPDIFF(
  SQL_TSI_DAY, 
  TIMESTAMPADD(
    SQL_TSI_YEAR, 
    -1, 
    TIMESTAMPADD(
      SQL_TSI_DAY, 
      1-DAYOFMONTH("Time"."T00 Calendar Date"), 
      "Time"."T00 Calendar Date"
    )
  ), 
  TIMESTAMPADD(
    SQL_TSI_DAY, 
    -DAYOFMONTH("Time"."T00 Calendar Date"), 
    "Time"."T00 Calendar Date"
  )
) + 1

The formatting is just to make it readable, it can be a single line. You can easily replace the "Time"."T00 Calendar Date"in the formula by CURRENT_DATEif you want to always make the calculation based on the current date, or you can also use any other date column of your dataset in it.

Give it a read, make sure you understand what's going on in the formula (because that's what let you adjust it if needed), and you are good to go.

It isn't going to be the best ever formula to be in a dataset, if you can calculate that in your database and add it somewhere in a dataset it will be better from a performance point of view. But it does the job for sure.