Sql query performance tuning

Question

Hi all,
I am facing problem to extract data and insert into another table. I need to fetch part number from all_parts_already and then based on the result I need to extract place from all_parts and insert into all_parts_already.
Sample data in the tables:
table all_parts_already:
part part_desc technique company place
1 A Engine TVS B1
1 Av Engine TVS B2
1 Ab Engine TVS B3
2 Ah Engine TVS B3
2 Ap Engine TVS B2
table all_parts:
technique company place
Engine TVS B1
Kim TVS B2
Engine TVS B3
Engine TVS B4
XXXXX TVS B5
Engine TVS B6
for c1 in (select distinct parts from all_parts_already where
technique = 'Engine' and
Company = 'TVS' ) loop
for c2 in (select distinct place from all_parts where
technique = 'Engine' and
Company ='TVS' and
minus
select distinct Place from all_parts_already where
technique = 'Engine' and
Company = 'TVS' and
parts = c1.parts ) loop
insert into all_parts_already (select c2.place,place_desc,c1.parts,c2.place from place_master where parts=c1.parts and place=c2.place);
end loop;
end loop;
the data i am dealing with is in millions. One technique may have 1000 parts. One part may have 500 places. So the loop runs that many times creating the delay.
Please tell me how to move forward.I am getting the output i need but the time it takes is too much(goes on to days)
Thanks a lot

Answer 1

RESULT TABLE AS BELOW
part place machine country
P2 C3 M1 I1
P2 C4 M1 I1
P4 C1 M1 I1
P4 C2 M1 I1
P4 C4 M1 I1
P3 C1 M1 I1
P3 C2 M1 I1
P3 C3 M1 I1I don't get the relationship to country.
How do you determine what the missing country is?
You say on all of the above the I1 is missing and yet in table 'a' parts can have different countries?
That point aside....
This probably isn't the simplest way to do it, but maybe this is moving in the right direction:
SQL> with b as
  2  (select 'c1' place, 'm1' machine from dual union all
  3   select 'c2' place, 'm1' machine from dual union all
  4   select 'c3' place, 'm1' machine from dual union all
  5   select 'c4' place, 'm1' machine from dual)
  6  ,    a as
  7  (select 'p1' part, 'c1' place, 'm1' machine, 'i1' country from dual union all
  8   select 'p1' part, 'c2' place, 'm1' machine, 'i2' country from dual union all
  9   select 'p1' part, 'c3' place, 'm1' machine, 'i3' country from dual union all
10   select 'p1' part, 'c4' place, 'm1' machine, 'i1' country from dual union all
11   select 'p2' part, 'c1' place, 'm1' machine, 'i1' country from dual union all
12   select 'p2' part, 'c2' place, 'm1' machine, 'i2' country from dual union all
13   select 'p4' part, 'c3' place, 'm1' machine, 'i3' country from dual union all
14   select 'p3' part, 'c4' place, 'm1' machine, 'i1' country from dual)
15  ,    x as
16  (select distinct part, machine
17   from   a)
18  ,   all_parts as
19  (select b.place, x.part, x.machine
20   from   x
21   ,      b
22   where  b.machine = x.machine)
23  select *
24  from   all_parts ap
25  ,      a         aa
26  where  aa.place   (+) = ap.place
27  and    aa.part    (+) = ap.part
28  and    aa.machine (+) = ap.machine
29  --and    aa.place is null
30  ;
PL PA MA PA PL MA CO
c1 p1 m1 p1 c1 m1 i1
c2 p1 m1 p1 c2 m1 i2
c3 p1 m1 p1 c3 m1 i3
c4 p1 m1 p1 c4 m1 i1
c1 p2 m1 p2 c1 m1 i1
c2 p2 m1 p2 c2 m1 i2
c3 p4 m1 p4 c3 m1 i3
c4 p3 m1 p3 c4 m1 i1
c2 p4 m1
c3 p2 m1
c3 p3 m1
c2 p3 m1
c4 p4 m1
c4 p2 m1
c1 p4 m1
c1 p3 m1
16 rows selected.
SQL> Just uncomment out the line "and aa.place is null" to get the missing rows.
Edited by: Dom Brooks on Nov 10, 2011 12:09 PM