1)After updating on complex view by using instead of trigger which table will affect first and why.
The trigger takes control of the ENTIRE ddl. So it can anything that *you* specify.
For example, here's one that does NOTHING with the original data ...
SQL> drop table t purge;
Table dropped.
SQL> create table t as
2 select mod(rownum,100) x, rownum y
3 from dual
4 connect by level <= 100000;
Table created.
SQL>
SQL> create or replace
2 view V as
3 select x, sum(y) tot_y
4 from t
5 group by x;
View created.
SQL>
SQL>
SQL> drop table t1 purge;
Table dropped.
SQL> create table t1 ( a date, b int );
Table created.
SQL>
SQL> create or replace
2 trigger trg
3 instead of insert or update or delete
4 on v
5 begin
6 insert into t1 values (sysdate,1);
7 end;
8 /
Trigger created.
SQL>
SQL> insert into v values (1,2);
1 row created.
SQL> select * from t1;
A B
--------- ----------
03-JUL-16 1
1 row selected.
SQL> delete from v;
100 rows deleted.
SQL> select * from t1;
A B
--------- ----------
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
03-JUL-16 1
101 rows selected.
SQL>
So they can be very dangerous in terms of not providing a "sensible" result if you are not careful.