Self join sub query help

Question

This is my table
Doctor     Appointment Date     Rating
111     23/12/2009     G1
111     23/12/2009     G2
111     23/12/2009     B
111          23/12/2009 D
111     23/12/2008     G1
111          23/12/2008 B
111     23/12/2007     G2
111          23/12/2007 B
111     23/12/2006     D
For Doctor 111, on a given appointment date he has multiple ratings.
I have to fetch the rows (with doctor id, appointment date)
I) which does not have G1 rating but has G2 rating for each appointment date
And vice versa ( does not have G2 rating but has G1 rating for each appointment date)
And
II) if the particular appointment date has both G1 and G2 rating then I can ignore that appointment date – i.e do not fetch any rows if that appointment date has both G1 and G2 ratings.
And
III)     if the appointment date has no G1 or G2 rating, then display the appointment date and the doctor.
So, my result set should have the output from all the three conditions for each doctor.
I am not sure how it works. I am trying using correlated subqueries any suggestions???

Answer 1

ME_XE?with data as
  2  (
  3     select 111 as doc_id, to_date('23/12/2009', 'dd/mm/yyyy') as app_date, 'G1'   as app_type from dual union all
  4     select 111 as doc_id, to_date('23/12/2009', 'dd/mm/yyyy') as app_date, 'G2'   as app_type from dual union all
  5     select 111 as doc_id, to_date('23/12/2009', 'dd/mm/yyyy') as app_date, 'B'    as app_type from dual union all
  6     select 111 as doc_id, to_date('23/12/2009', 'dd/mm/yyyy') as app_date, 'D'    as app_type from dual union all
  7     select 111 as doc_id, to_date('23/12/2008', 'dd/mm/yyyy') as app_date, 'G1'   as app_type from dual union all
  8     select 111 as doc_id, to_date('23/12/2008', 'dd/mm/yyyy') as app_date, 'B'    as app_type from dual union all
  9     select 111 as doc_id, to_date('23/12/2007', 'dd/mm/yyyy') as app_date, 'G2'   as app_type from dual union all
10     select 111 as doc_id, to_date('23/12/2007', 'dd/mm/yyyy') as app_date, 'B'    as app_type from dual union all
11     select 111 as doc_id, to_date('23/12/2006', 'dd/mm/yyyy') as app_date, 'D'    as app_type from dual
12  )
13  select
14     doc_id,
15     app_date,
16     app_type
17  from
18  (
19     select
20        sum( case when app_type = 'G1' then 1 else 0 end ) over (partition by doc_id, app_date) as g1,
21        sum( case when app_type = 'G2' then 1 else 0 end ) over (partition by doc_id, app_date) as g2,
22        doc_id,
23        app_date,
24        app_type
25     from data
26  )
27  where g1 = 0 and g2 = 0
28  or
29  (
30     g1 > 0 and g2 = 0
31  )
32  or
33  (
34     g2 > 0 and g1 = 0
35  );
            DOC_ID APP_DATE                   APP_TY
               111 23-DEC-2006 12 00:00       D
               111 23-DEC-2007 12 00:00       B
               111 23-DEC-2007 12 00:00       G2
               111 23-DEC-2008 12 00:00       G1
               111 23-DEC-2008 12 00:00       B
5 rows selected.
Elapsed: 00:00:00.59
ME_XE?
ME_XE?select * from v$version;
BANNER
Oracle Database 10g Express Edition Release 10.2.0.1.0 - Product
PL/SQL Release 10.2.0.1.0 - Production
CORE    10.2.0.1.0      Production
TNS for Linux: Version 10.2.0.1.0 - Production
NLSRTL Version 10.2.0.1.0 - Production
5 rows selected.
Elapsed: 00:00:00.64Is one method using analytics (which allow you to perform operations over a 'window' of data, which seems to be what you need to do).