Update query in procedure

Question

I have 2 tables aas below.
create table main_data(id number,all_set_flag char(1),set_time timestamp)
create table sub_data(id number,set_time timestamp);
insert into main_data values(1,'','');
insert into main_data values(1,'','');
insert into main_data values(1,'','');
insert into main_data values(1,'','');
insert into main_data values(2,'Y','');
insert into main_data values(2,'Y','');
insert into main_data values(2,'Y','');
My requirement is 1)For a particular id if all the all_set_flag is 'Y' then I need to insert the distinct id record and set_time as systimestamp in sub_data table
2)Afer inserting into subdata for that particular id I need to update back in main_data set_time same as set_time in sub_data.
a)For the part -1 I wrote an insert statement like this.For part-2 how can I update it bk?.
b)I am going to use the below query in a procedure.Is there any better way to achieve this than below approach.Please note that my main_data have more columns(200 columns)
and there are millions of records.10 million atleast.Please suggest.
insert into sub_data(id,set_time)
(select
distinct
id,
systimestamp
from
select
id,
sum(case when all_set_flag = 'Y' then 1 else 0 end) over (partition by id) as match_cnt,
count(*) over (partition by id) as total_count
from main_data where not exists (Select null from sub_data where sub_data.id = main_data.id
where match_cnt = total_count);
select * from sub_data

Answer 1

I am using a PL/SQL Collection type and doing it. Was think of doing it in plain SQL but combaining the Insert and Update i guess needs some PL/SQL.
Here is what i did
declare
  type main_data_id is table of main_data.id%type;
  type main_data_tm is table of main_data.set_time%type;
  lMainDataId main_data_id;
  lmainDataTm main_data_tm;
begin
  select distinct id, systimestamp
  bulk collect into lMainDataId, lMainDataTm
    from (
          select a.* , count(*) over(partition by id) id_cnt, count(decode(all_set_flag, 'Y',1,null)) over(partition by id) flag_cnt
            from main_data a
   where id_cnt = flag_cnt;
  forall i in 1..lMainDataId.count
    insert into sub_data(id, set_time) values (lMainDataId(i), lMainDataTm(i));
  forall i in 1..lMainDataId.count
    update main_data set set_time = lMainDataTm(i) where id = lMainDataId(i);
  commit;
end;
/Waiting to see if someone has a better approach to it :)
Note: I have not used LIMIT clause in BULK COLLECT because i dont have idea of the volume of data involved. If there is large volume of data then you need to consider using it.