Repeat previous values by date in hive

Use a cross join to generate the rows. Then a left join combine the two tables. And finally use last_value() to bring in missing values:

select d.date, i.id,
       last_value(amount, TRUE) over (partition by id order by d.date) as amount
from date d cross join
     (select distinct id from data) i left join
     data t
     on d.date = t.date and i.id = t.id;

The last_value() with the second argument ignores NULL values, so it will "go back" to get the most recent non-NULL value.

First outer join two tables, then following

with join_table as (
select coleasce(t1.date, t2.date) as date, t2.id, t2.amount
  from table1 as t1
  outer join table2 as t2
    on t1.date = t2.date
)
, rank_table as ( 
select *, SUM(id) OVER (ORDER BY d_day ROWS BETWEEN unbounded preceding and CURRENT ROW) as id_rnk,
SUM(amount) OVER (ORDER BY d_day ROWS BETWEEN unbounded preceding and CURRENT ROW) as amount_rnk
  from join_table
)
select *, max(id) over (partition by id_rnk), max(amount) over (partition by amount_rnk)
  from rank_table

Check this answer for more details.