SergeySafaryanc

select distinct
	id, 
	case 
		when
			lag(dr) over(partition by id order by rn) is null	
			or lag(dr) over(partition by id order by rn) = dr
		then 
			min(date_from) over(partition by id, dr order by rn)
		else
			min(date_from) over(partition by id, dr, rn order by rn)
	end as date_from,
	case 
		when
			lead(dr) over(partition by id order by rn) is null	
			or lead(dr) over(partition by id order by rn) = dr
		then 
			max(date_to) over(partition by id, dr order by rn DESC)
		else
			max(date_to) over(partition by id, dr, rn order by rn DESC)
	end as date_to,
	param1,
	param2
from  (
	select
		id, date_from, date_to, param1, param2,
		row_number() over(partition by id order by date_from ASC) as rn,
		dense_rank() over(partition by id order by param1, param2) as dr,
	from
		Table
	) a1