Pivoting; dynamic pivot_in_clause

Question

I have this situation:
create table pivot_data
(group_id NUMBER(10),
address_state VARCHAR2(20),
address_zip VARCHAR2(5),
skill_id NUMBER(10),
person_id NUMBER(10),
person_name VARCHAR2(20)
INSERT INTO pivot_data VALUES (1,'VIRGINIA','22222',10,100,'JOHN');
INSERT INTO pivot_data VALUES (1,'VIRGINIA','22222',10,200,'JEFF');
INSERT INTO pivot_data VALUES (1,'VIRGINIA','22222',10,300,'GEORGE');
INSERT INTO pivot_data VALUES (2,'VIRGINIA','22222',10,400,'KATE');
INSERT INTO pivot_data VALUES (2,'VIRGINIA','22222',10,500,'KELLY');
INSERT INTO pivot_data VALUES (3,'VIRGINIA','22222',10,600,'SAM');
INSERT INTO pivot_data VALUES (4,'VIRGINIA','22222',20,100,'JOHN');
INSERT INTO pivot_data VALUES (4,'VIRGINIA','22222',20,200,'JEFF');
INSERT INTO pivot_data VALUES (4,'VIRGINIA','22222',20,300,'GEORGE');
INSERT INTO pivot_data VALUES (5,'VIRGINIA','22222',20,400,'KATE');
INSERT INTO pivot_data VALUES (5,'VIRGINIA','22222',20,500,'KELLY');
INSERT INTO pivot_data VALUES (5,'VIRGINIA','22222',20,600,'SAM');
COMMIT;
SELECT * FROM pivot_data;
1     VIRGINIA     22222     10     100     JOHN
1     VIRGINIA     22222     10     200     JEFF
1     VIRGINIA     22222     10     300     GEORGE
2     VIRGINIA     22222     10     400     KATE
2     VIRGINIA     22222     10     500     KELLY
3     VIRGINIA     22222     10     600     SAM
4     VIRGINIA     22222     20     100     JOHN
4     VIRGINIA     22222     20     200     JEFF
4     VIRGINIA     22222     20     300     GEORGE
5     VIRGINIA     22222     20     400     KATE
5     VIRGINIA     22222     20     500     KELLY
5     VIRGINIA     22222     20     600     SAM
Business rules:
1.People from a given state, zip code are grouped based on their skillset.
2.A person may have more than one skillset
3.A person can be part of different groups . Even a person can belong to more than one group for same skill id in case
groups are made again for same skill set
4.A group can have at the most 3 members
5.I am using Oracle 11g. Don't know how to use 11g pivot function because pivot_in_clause in my case is dynamic
I want to pivot the data in order to show each group in one row:
state       zip     skill_id member1  member1_name member2  member2_name   member3  member3_name
VIRGINIA    22222   10       100      JOHN          200     JEFF            300       GEORGE
VIRGINIA    22222   10       400      KATE          500     KELLY         
VIRGINIA    22222   10       600      SAM          
VIRGINIA    22222   20       100      JOHN          200     JEFF            300       GEORGE
VIRGINIA    22222   20       400      KATE          500     KELLY           600       SAM
Thanks for the help!
RN

Answer 1

Hi,
Thanks for posting the CREATE TABLE and INSERT statements; that's very helpful!
RN wrote:
... Business rules:
1.People from a given state, zip code are grouped based on their skillset.
2.A person may have more than one skillset
3.A person can be part of different groups . Even a person can belong to more than one group for same skill id in case
groups are made again for same skill set
4.A group can have at the most 3 members
5.I am using Oracle 11g. Don't know how to use 11g pivot function because pivot_in_clause in my case is dynamic
It's dynamic only in that the value that tells which person appears in which column has to be derived at run time.
It's not dynamic in the sense that the number of columns is fixed: you always want 9 columns (3 for group information, and then 3 pivoted sets of 2 columns each).
I want to pivot the data in order to show each group in one row:
state       zip     skill_id member1  member1_name member2  member2_name   member3  member3_name
VIRGINIA    22222   10       100      JOHN          200     JEFF            300       GEORGE
VIRGINIA    22222   10       400      KATE          500     KELLY         
VIRGINIA    22222   10       600      SAM          
VIRGINIA    22222   20       100      JOHN          200     JEFF            300       GEORGE
VIRGINIA    22222   20       400      KATE          500     KELLY           600       SAM Sorry, I'm not at an Oracle 11 database right now, so I can't test a solution using SELECT ... PIVOT. I'm not sure it would really help, anyway.
Here's how you can do it in Oracle 9 (and up):
WITH     got_c_num     AS
     SELECT     group_id, address_state
     ,      address_zip, skill_id
     ,      person_id, person_name
     ,     ROW_NUMBER () OVER ( PARTITION BY  group_id, address_state
                               ,                     address_zip, skill_id
                         ORDER BY        person_id
                       )         AS c_num
     FROM    pivot_data
SELECT       address_state
,        address_zip, skill_id
,        MIN (CASE WHEN c_num = 1 THEN person_id   END)     AS member1
,        MIN (CASE WHEN c_num = 1 THEN person_name END)     AS member1_name
,        MIN (CASE WHEN c_num = 2 THEN person_id   END)     AS member2
,        MIN (CASE WHEN c_num = 2 THEN person_name END)     AS member2_name
,        MIN (CASE WHEN c_num = 3 THEN person_id   END)     AS member3
,        MIN (CASE WHEN c_num = 3 THEN person_name END)     AS member3_name
FROM       got_c_num
GROUP BY  group_id, address_state
,            address_zip, skill_id
ORDER BY  group_id, address_state
,            address_zip, skill_id
;