Trying to simplify multiple steps into one - please help

Question

I have a set up of steps to perform on 3 tables in order to populate two tables, please suggest best way (performance wise) to do this
CREATE TABLE AE
  EID    VARCHAR2(100 BYTE),
  UKEY   VARCHAR2(1000 BYTE),
  TNAME  VARCHAR2(10 BYTE)
CREATE TABLE TAB1V
  PID    VARCHAR2(100 BYTE),
  PUKEY  VARCHAR2(100 BYTE)
(join ae and tab1v with this condition: ae.ukey = tab1v.pukey);
CREATE TABLE PES
  EID    VARCHAR2(100 BYTE),
  UKEY   VARCHAR2(100 BYTE),
  TNAME  VARCHAR2(1000 BYTE)
(join condition with other tables, use ukey)
insert into ae (eid,ukey,tname) values ('eid1','ukey123','tname123');
insert into ae (eid,ukey,tname) values ('eid2','ukey123','tname123');
insert into ae (eid,ukey,tname) values ('eid3','ukey222','tname222');
insert into ae (eid,ukey,tname) values ('eid4','ukey222','tname222');
insert into ae (eid,ukey,tname) values ('eid5','ukey456','tname456');
insert into ae (eid,ukey,tname) values ('eid6','ukey789','tname789');
insert into ae (eid,ukey,tname) values ('eid7','ukey789','tname7891');
insert into pes(eid,ukey,tname) values('eid2','ukey123','tname123');
insert into pes(eid,ukey,tname) values('eid6','ukey789','tname789');
insert into pes(eid,ukey,tname) values('eid5','ukey456','tname456');
insert into pes(eid,ukey,tname) values('eid2','ukey123','tname123');
insert into tab1v(pid,pukey) values ('pid1','ukey123');
insert into tab1v(pid,pukey) values ('pid2','ukey222');
insert into tab1v(pid,pukey) values ('pid3','ukey111');
insert into tab1v(pid,pukey) values ('pid4','ukey456');
insert into tab1v(pid,pukey) values ('pid5','ukey789');my goal:
i have to populate two tables
1. result_table
CREATE TABLE RESULT_TABLE
PID VARCHAR2(100 BYTE),
UKEY VARCHAR2(100 BYTE),
EID VARCHAR2(100 BYTE),
TNAME VARCHAR2(100 BYTE),
SRC VARCHAR2(4 BYTE)
2. no_match table
CREATE TABLE NO_MATCH
PID VARCHAR2(100 BYTE),
UKEY VARCHAR2(100 BYTE),
my flow of logic:
1. check all the rows that are common to both tab1v and ae table (by joining with ukey),
if there are ukeys that are present in tab1v table, but absent in ae table, populate
them in no_match table, I got this with below query
select pid,pukey from tab1v
where not exists
(select ukey from ae
where ae.ukey = tab1v.pukey);
result of no_match table should be:
pid3 ukey111
for all those rows that match, probably populate in a staging table
CREATE TABLE STG_TABLE1
PID VARCHAR2(20 BYTE),
UKEY VARCHAR2(100 BYTE),
TNAME VARCHAR2(100 BYTE),
EID VARCHAR2(100 BYTE)
insert into stg_table1(pid,ukey,eid,tname)
select a.pid,b.ukey,b.eid,b.tname
from tab1v a, ae b
where A.PUKEY = b.ukey
2.After doing point 1, So if I find common rows in both tab1v and ae table (present in stg_tab1), pick up those records and check if they are present in pes table ?
join stg_table1 and pes table with ukey again
3. populate records found / and not found in point 2, do the following check and populate result_table
for records found in both stg_table1 (in other words, present in both ae and tab1v) and pes, do the following check
if each ukey has multiple same tnames attached to it (ukey123 has tname123, twice ), src should be A
if there is unique tname per ukey (ukey456 has only tname456), src should be A
if each ukey has multiple different tnames (ukey 789 has tname789 and tname7891), they should be populated and src column should be B
result_table should be like below :
PID     UKEY     EID     TNAME SRC
pid1     ukey123     eid1     tname123 A
pid1     ukey123     eid2     tname123 A
pid4     ukey456     eid5     tname456 A
pid5     ukey789     eid6     tname789       B
pid5     ukey789     eid7     tname7891      B
pid2     ukey222     null     tname222 A
pid2     ukey222     null     tname222 A
for records found in stg_table1 but absent in pes, populate the above table and same following check, but assign null to eid : (last 2 rows in above data)
following can be used to find difference
select pid,ukey,tname from stg_table1
where not exists
(select ukey from pes
where pes.ukey = stg_table1.ukey);
if each ukey has multiple same tnames attached to it (ukey222 has tname222, twice ), src should be A
if there is unique tname per ukey src should be A
if each ukey has multiple different tnames they should be populated and src should be B (none in my example data)
I am getting the results in individual sql statements, but how do I populate result_table and no_match with above data, given the above source data with minimal sql statements and considering performace as a key factor ?

Answer 1

Why so many unanswered questions? 66 (47 unresolved) . Are they really unanswered? Please review your 47 unresolved questions and give credit to the people with either HELPFUL or ANSWERED as appropriate. They will appreciate it.
Please provide you 4 digit Oracle version number.
In 11g you could use the INSERT ALL multi-table insert to do this: http://docs.oracle.com/cd/B28359_01/server.111/b28286/statements_9014.htm
See the above doc at 'The next examples insert into multiple tables' where you will find this excerpt:
>
INSERT ALL
WHEN order_total < 1000000 THEN
INTO small_orders
WHEN order_total > 1000000 AND order_total < 2000000 THEN
INTO medium_orders
WHEN order_total > 2000000 THEN
INTO large_orders
SELECT order_id, order_total, sales_rep_id, customer_id
FROM orders;
>
This will let you insert into both tables with one query. Your query can use an OUTER JOIN between your tables so that the key value will be NULL when a record is not found. Then your WHEN condition will check the key value and if NULL will insert into one table and if NOT NULL will insert into the other.
Should not be a need for a staging table. Try to modify the example in the doc for your tables.