Inner join need help

Question

hi
I have following data
create table test1 (ind int,idd varchar(20), sec int, amt float)
create table test2 (ind int, id1 varchar(10), id2 varchar(10), sec int, qty float)
insert into test1 values (11, '1aa',1,100);
insert into test1 values (12, '1aa',1,200);
insert into test1 values (13, '2bb',2,500);
insert into test1 values ( 14,'2bb',2,600);
insert into test1 values ( 15, '3cc',3,100);
insert into test1 values ( 16, '4dd',4,100);
insert into test1 values ( 17, '1aa',5,5100);
insert into test1 values ( 18, '1aa',6,100);
insert into test2 values( 1, '1','aa',1, 300);
insert into test2 values( 2, '1','aa',1, 300);
insert into test2 values( 3, '2','bb',2, 700);
insert into test2 values( 4, '2','bb',2, 100);
insert into test2 values( 5, '3','cc',3, 400);
insert into test2 values( 6, '1','aa',5, 3100);
insert into test2 values( 7, '1','aa',6, 7100);
select test2.ind, idd, id1||id2 , test1.sec, amt, qty from test1 inner join test2 on idd=id1||id2 and test1.sec=test2.sec
order by indI am getting 11 records
1     1aa     1aa     1     100     300
     1     1aa     1aa     1     200     300
     2     1aa     1aa     1     100     300
     2     1aa     1aa     1     200     300
     3     2bb     2bb     2     500     700
     3     2bb     2bb     2     600     700
     4     2bb     2bb     2     600     100
     4     2bb     2bb     2     500     100
     5     3cc     3cc     3     100     400
     6     1aa     1aa     5     5100     3100
     7     1aa     1aa     6     100     7100
and following is my desired output
     1     1aa     1aa     1     100     300
     2     1aa     1aa     1     200     300
     3     2bb     2bb     2     500     700
     4     2bb     2bb     2     600     100
     5     3cc     3cc     3     100     400
     6     1aa     1aa     5     5100     3100
     7     1aa     1aa     6     100     7100
please help

Answer 1

Hi,
Thanks for posting the CREATE TABLE and INSERT statements! That's very helpful.
It's also helpful if you explain how you get the results you want from that data.
Why do you want 7 rows of output, not 11?
Given that you do want 7 rows of output, why do you want the results you posted, and not
IND  IDD  ID1_ID2  SEC  AMT  QTY
1    1aa  1aa        1     200  300
2    1aa  1aa        1     200  300
...or
IND  IDD  ID1_ID2  SEC  AMT  QTY
1    1aa  1aa         1     100  300
2    1aa  1aa        1     100  300
...or some other combination? Don't force people to spend time guessing, and don't give them a chance to guess wrong.
It looks like you're getting the output you want now, except when there are multiple rows with the same idd and sec in test1 (or the same id1, id2 and sec in test2). These are the join columns.
It looks like you only want one row of output for each gropup that has the same join columns.
Within each of those groups, do you only want
the row with the lowest test1.ind joined to the row with the lowest test2.ind,
the row with the 2nd lowest test1.ind joined to the row with the 2nd lowest test2.ind,
the row with the 3rd lowest test1.ind joined to the row with the 3rd lowest test2.ind,
and so on?
If so, use the analytic ROW_NUMBER functions to indicate whihc i the lowest, 2nd lowest, 3rd lowest, and so on:
WITH   test1_with_r_num  AS
     SELECT     ind, idd, sec, amt
     ,     ROW_NUMBER () OVER ( PARTITION BY  idd, sec
                               ORDER BY          ind
                       )         AS r_num
     FROM    test1
,     test2_with_r_num  AS
     SELECT     ind, id1, id2, sec, qty
     ,     ROW_NUMBER () OVER ( PARTITION BY  id1, id2, sec
                               ORDER BY          ind
                       )         AS r_num
     FROM     test2
select    t2.ind
,        t1.idd
,        t2.id1 || t2.id2     AS id1_id2
,       t1.sec
,        t1.amt
,       t2.qty
from                 test1_with_r_num     t1
inner join          test2_with_r_num      t2     on     t1.idd       = t2.id1 || t2.id2
                                        and     t1.sec       = t2.sec
                              and     t1.r_num  = t2.r_num
order by  t2.ind
;This happens to give the results you requested from the data you posted. It may be purely by coincidence.
What if there are an unequal number of rows with the same join conditions in the two tables?
For example, what if we add another row to test1:
insert into test1 (ind, idd, sec, amt) values (91, '1aa',1,125);but don't add any new rows to test2?
What if we add a row only to test2
insert into test2 (ind, id1, id2, sec, ity) values( 9, '1','aa',6, 7199);? What results would you want in these cases?