hard.sql

--drop table if exists #seq
--;with t(date,status) as (
--    select convert(date,dateadd(d,row_number() over (order by @@dbts),{d'2018-12-31'}))
--          ,abs(sign(BINARY_CHECKSUM(newid())%2))
--from string_split(REPLICATE('1,',30),',') a
--)
--select * into #seq from t


select s1.date
	  ,s1.status
	  ,datediff(d,isnull(last_one,min_d),s1.date) successive_days	
from #seq s1
outer apply (
    select top 1 s2.date
    from #seq s2
    where s1.date >=s2.date
        and s2.status =  1
    order by s2.date desc) f(last_one)
cross join (
    select dateadd(d,-1,min(date))
    from #seq) q(min_d)

select date
      ,status
      ,(1^status) * row_number() over (partition by pkey order by date)
from (
    select date
          ,status
          ,sum(change_event) over (order by date) pkey
    from (
        select date
              ,status
              ,1 + 0 * nullif(s1.status,lag(s1.status) over (order by date)) change_event
        from #seq s1
    ) q
) q
order by date