What is the best method?

Question

Hi,
Please find the queries below;
SQL> create table aat1(key1 number(5), verno number(5),text1 varchar2(100),text2
varchar2(100))
  2  /
Table created.
SQL> alter table aat1 add constraint aap1 primary key (key1,verno)
  2  /
Table altered.
SQL> alter table aat1 add constraint aau1 unique (verno,text1)
  2  /
Table altered.
SQL> select * from aat1;
      KEY1      VERNO       TEXT1   TEXT2
         1          0         ABC       DEF
         1          1         ABC       TTT
         2          0         ZZZ       MMM
SQL> create table aat2 as select * from aat1;
Table created.
SQL> alter table aat2 drop column verno;
Table altered.
SQL> alter table aat2 drop column key1;
Table altered.
SQL> update aat2 set text1 = 'XXX' where text1='ZZZ';
1 row updated.
SQL> delete aat2 where text2 = 'TTT';
1 row deleted.
SQL> update aat2 set text2 = 'IOU' where text2 = 'DEF';
1 row updated.
SQL> select * from aat2;
TEXT1            TEXT2
ABC                IOU
XXX                MMMNow what I want is;
1. Whatever record in both the tables having matching text1 value that should have a new record in aat1 with verno as 0 and the previous record/records' verno should be verno+1 keeping key1 value as same in aat1.
2. The text2 value of aat2 will be the updated value for text2 field of this new inserted record in aat1.
3. If any non matching text1 is there that should also be entered into aat1 where verno will be 0 and a new key1(could be taken from oracle sequence in practical scenario).
This is what I have done. I would need suggestion if this can be done in a more performance effective way.Basically any better way of doing this thing.
SQL> update aat1 a set a.verno = a.verno + 1 where a.text1 = (select distinct b.text1 from aat2 b where b.text1 = a.text1 )
  2  /
2 rows updated.
SQL> select * from aat1;
      KEY1      VERNO        TEXT1   TEXT2
         1          1        ABC       DEF
         1          2        ABC       TTT
         2          0        ZZZ       MMM
SQL> insert into aat1 select a.key1,0,b.text1,b.text2 from aat1 a,aat2 b where a.text1 = b.text1 and a.verno = 1 ;
1 row created.
SQL> select * from aat1;
      KEY1      VERNO TEXT1    TEXT2
         1          1        ABC       DEF
         1          2        ABC       TTT
         2          0        ZZZ       MMM
         1          0        ABC       IOU
SQL> MERGE INTO aat1 a
  2    USING aat2 b
  3      ON (a.text1 = b.text1)
  4    WHEN NOT MATCHED THEN
  5      INSERT (key1,verno,text1,text2)
  6      VALUES (3,0,b.text1,b.text2)
  7  /
1 row merged.
SQL> select * from aat1 order by key1,verno;
      KEY1      VERNO TEXT1   TEXT2
         1          0 ABC          IOU
         1          1 ABC          DEF
         1          2 ABC          TTT
         2          0 ZZZ          MMM
         3          0 XXX          MMM
SQL>Please help me.
regards,
Dipankar.

Answer 1

None other than what is already done.!
Just add this hint and check the performance/explain plan/tkprof.
MERGE /*+ parallel(table_name,10) append */
Nothing other than this.!
Ss