如何根据日期间隔获取每个 ID 和月份的最新事件列表？

Question

I have a table registering different event dates for unique identifiers - one row for each Id-event-date combination. Each unique identifier can have multiple dates, according to the number of events happening to that Id. There are no two events occurring on the same date for individual Ids.

First, I wanted to identify the unique Ids for which the latest event has occurred for more than 7 days to current date. This is my working approach:

WITH latest_event AS
(
SELECT Id,
       date,
       event,
       ROW_NUMBER() OVER(PARTITION BY Id ORDER BY date DESC) AS row_number
  FROM `table`
),

SELECT Id,
       date,
       event,
  FROM latest_stage 
 WHERE row_number = 1
   AND DATE_DIFF(CURRENT_DATE(),date, DAY) > 7
 ORDER BY date

Now I would like to obtain the same information, but on a monthly basis, instead of just current date. For instance for every month in 2021 and 2022, identify the number of unique Ids which, by the end of each month, did not have an event in the last 7 days. I am struggling to understand how to do this calculation for every month in these two years.

Answer 1

I am not sure this will work. Please improve it yourself.

with tbl as 
(Select cast(rand()*10 as int64) Id,  date,"E" event from unnest(generate_date_array("2021-01-01","2022-09-23",interval 1 day)) as date)
select *,
date_trunc(date,month)  as month_trunc,
lag(date) over last_date_window as last_date,
ifnull(date_diff(date, lag(date) over last_date_window,day),99) as diff,
max(date) over this_month_window as last_date_of_month
 from tbl
 qualify
  ifnull(date_diff(date, lag(date) over last_date_window,day),99)>7
  and date=max(date) over this_month_window
 window
  last_date_window as (partition by id order by date),
  this_month_window as (partition by date_trunc(date,month))

 
order  by date