Grouping based on Start and End Year

Question

Hello everyone,
Need your help in solving one SQL query
My data look like this
cid rid year title
52     1000     2001     1
52     1000     2002     1
52     1000     2003     6
52     1000     2004     6
52     1000     2005     6
52     1000     2007     1
52     1001     2003     1
52     1001     2004     1
52     1001     2005     1
52     1001     2006     3
52     1001     2007     3
output I am required to produce is
CID          RID          START-END     TITLE
52          1000     2001-2002     1
52          1000     2003-2005     6
52          1000     2007-2007     1
52          1001     2003-2005     1
52          1001     2006-2007     3
here is script for table and data
create table T (
cid NUMBER,
rid NUMBER,
start_year NUMBER,
title NUMBER
insert into t (cid, rid, start_year, title) values (52, 1000, 2001, 1);
insert into t (cid, rid, start_year, title) values (52, 1000, 2002, 1);
insert into t (cid, rid, start_year, title) values (52, 1001, 2003, 1);
insert into t (cid, rid, start_year, title) values (52, 1000, 2003, 6);
insert into t (cid, rid, start_year, title) values (52, 1001, 2004, 1);
insert into t (cid, rid, start_year, title) values (52, 1000, 2004, 6);
insert into t (cid, rid, start_year, title) values (52, 1001, 2005, 1);
insert into t (cid, rid, start_year, title) values (52, 1000, 2005, 6);
insert into t (cid, rid, start_year, title) values (52, 1001, 2006, 3);
insert into t (cid, rid, start_year, title) values (52, 1001, 2007, 3);
insert into t (cid, rid, start_year, title) values (52, 1000, 2007, 1);
commit;
I was able to go this far (some what close to what I need but not exactly what I need)
select cid, rid, min(year2), max(year1), title
from (
select tc.cid, tc.rid, tc.title, tc.start_year year1, tp.start_year year2, decode (tc.title, tp.title, 'N', 'Y') gap
from t tc,
t tp
where tc.start_year -1 = tp.start_year
and tc.cid = tp.cid
and tc.rid = tp.rid
group by cid, rid, title, gap
Thanks for considering
Naseer,
Houston-TX

Answer 1

I'm not sure how elegant this is, but the following tweak to your query solved the problem I posed above:
WITH      got_grp_id     AS
     SELECT     cid, rid, start_year, title
     ,     ROW_NUMBER () OVER ( PARTITION BY  cid, rid
                               ORDER BY          start_year
           -     ROW_NUMBER () OVER ( PARTITION BY  cid, rid, title
                             ORDER BY          start_year
              + NVL(LAG(start_year,1) OVER(PARTITION BY cid,rid,title
                                       ORDER BY cid,rid,title),start_year - 1
                                   ) + 1 - start_year        AS grp_id
     FROM    t
SELECT       cid
,       rid
,       MIN (start_year) || '-'
                        || MAX (start_year)     AS start_end
,       title
FROM       got_grp_id
GROUP BY  cid, rid, title, grp_id
ORDER BY  cid, rid, start_end
CID  RID  START_END  TITLE 
52  1000  2001-2002  1 
52  1000  2007-2007  1 
52  1001  2003-2005  1 
52  1001  2006-2007  3