I have tried any number of permutations trying to monitor null percentage on a column, but every time the SQL it generates is:

with nothing as (select 1 as num)
    select * from nothing where num = 2

my most recent attempt looks like this:

version: 2

models:
  - name: analytic_events_prep
    description: '{{ doc("analytic_events_prep") }}'
    config: 
      elementary:
        timestamp_column: 'timestamp'
    columns:
      - name: account_id
        tests:
          - elementary.column_anomalies:
              column_anomalies: 
                - null_percent
              anomaly_direction: spike
              training_period:
                period: day
                count: 30
              detection_period:
                period: day
                count: 30
              seasonality: day_of_week
              time_bucket: # Daily by default
                period: month
                count: 1

I am using the most recent version of the package:

- package: elementary-data/elementary
    version: 0.16.1

Does anyone know what I am missing?

Hi @Kevin Burnham, What do you see in the report? Does it give you a warning that there is not enough data to calculate the score? If you use a time_bucket of one month but have a detection_period of 30 days it means it only calculates one score. Can you try and change the time_bucket to:

time_bucket:
  period: day
  count: 1