Showing posts with label valuefor. Show all posts
Showing posts with label valuefor. Show all posts

Wednesday, March 28, 2012

Finding Consecutive Values

I'm stuck in trying to find out how to solve this.
Within a record set, I need to know when a id has been a certain value
for x consecutive years. For example,
Create table #Test (
tid int NOT NULL,
yr int NOT NULL,
value int NOT NULL)
insert into #Test values (100, 1998, 0)
insert into #Test values (100, 1999, 0)
insert into #Test values (100, 2000, 0)
insert into #Test values (100, 2001, 0)
insert into #Test values (100, 2002, 1)
insert into #Test values (100, 2003, 0)
insert into #Test values (100, 2004, 0)
insert into #Test values (100, 2005, 0)
insert into #Test values (100, 2006, 0)
insert into #Test values (100, 2007, 0)
insert into #Test values (100, 2008, 1)
insert into #Test values (200, 1999, 0)
insert into #Test values (200, 2001, 0)
insert into #Test values (200, 2002, 0)
insert into #Test values (300, 2001, 0)
insert into #Test values (300, 2002, 0)
insert into #Test values (300, 2003, 0)
insert into #Test values (300, 2004, 0)
insert into #Test values (300, 2005, 0)
FIND: tid, startdate, enddate when value is 0 for 4 or more consecutive
months.
RESULTS would yield
tid startdate enddate years
100 1998 2001 4
100 2003 2007 5
300 2001 2005 5
TIAselect t_from.tid, t_from.yr start_year, t_to.yr end_year, t_from.value
from
(select tid, yr, value from Test t1 where not exists(
select 1 from Test t2 where t1.tid=t2.tid and t1.value=t2.value and
t1.yr=(t2.yr+1))
)t_from
join
(select tid, yr, value from Test t1 where not exists(
select 1 from Test t2 where t1.tid=t2.tid and t1.value=t2.value and
t1.yr=(t2.yr-1))
)t_to
on t_from.tid=t_to.tid and t_from.value=t_to.value and
(t_from.yr+3)<=t_to.yr
and t_from.yr + (select count(*) from Test t where t_from.tid=t.tid
and t_from.yr<t.yr and t.yr<t_to.yr and t_from.value=t.value) + 1 =
t_to.yr
order by t_from.tid, t_from.yr
tid start_year end_year value
-- -- -- --
100 1998 2001 0
100 2003 2007 0
300 2001 2005 0
(3 row(s) affected)|||select tid, min(yr) as start, maxyr as [end], count(*) as yrs
from (
select *,
isnull(
(select max(yr)
from #test t2
where t2.tid=t1.tid
and t2.yr>=t1.yr
and t2.value=0
and t2.yr<=(select min(yr) from #test where tid=t2.tid and
yr>=t1.yr and value=1)
),
(select max(yr)
from #test t3
where t3.tid=t1.tid
)
) as maxyr
from #test t1
where value=0
) t4
group by tid, maxyr
having count(*)>=4
order by tid, start
carmaboy@.gmail.com wrote:
> I'm stuck in trying to find out how to solve this.
> Within a record set, I need to know when a id has been a certain value
> for x consecutive years. For example,
> Create table #Test (
> tid int NOT NULL,
> yr int NOT NULL,
> value int NOT NULL)
> insert into #Test values (100, 1998, 0)
> insert into #Test values (100, 1999, 0)
> insert into #Test values (100, 2000, 0)
> insert into #Test values (100, 2001, 0)
> insert into #Test values (100, 2002, 1)
> insert into #Test values (100, 2003, 0)
> insert into #Test values (100, 2004, 0)
> insert into #Test values (100, 2005, 0)
> insert into #Test values (100, 2006, 0)
> insert into #Test values (100, 2007, 0)
> insert into #Test values (100, 2008, 1)
> insert into #Test values (200, 1999, 0)
> insert into #Test values (200, 2001, 0)
> insert into #Test values (200, 2002, 0)
> insert into #Test values (300, 2001, 0)
> insert into #Test values (300, 2002, 0)
> insert into #Test values (300, 2003, 0)
> insert into #Test values (300, 2004, 0)
> insert into #Test values (300, 2005, 0)
>
> FIND: tid, startdate, enddate when value is 0 for 4 or more consecutive
> months.
> RESULTS would yield
> tid startdate enddate years
> 100 1998 2001 4
> 100 2003 2007 5
> 300 2001 2005 5
> TIA
>|||A different, shorter version:
DROP TABLE Foobar;
CREATE TABLE Foobar
(tid INTEGER NOT NULL,
yr INTEGER NOT NULL,
value INTEGER NOT NULL);
INSERT INTO Foobar VALUES (100, 1998, 0);
INSERT INTO Foobar VALUES (100, 1999, 0);
INSERT INTO Foobar VALUES (100, 2000, 0);
INSERT INTO Foobar VALUES (100, 2001, 0);
INSERT INTO Foobar VALUES (100, 2002, 1);
INSERT INTO Foobar VALUES (100, 2003, 0);
INSERT INTO Foobar VALUES (100, 2004, 0);
INSERT INTO Foobar VALUES (100, 2005, 0);
INSERT INTO Foobar VALUES (100, 2006, 0);
INSERT INTO Foobar VALUES (100, 2007, 0);
INSERT INTO Foobar VALUES (100, 2008, 1);
INSERT INTO Foobar VALUES (200, 1999, 0);
INSERT INTO Foobar VALUES (200, 2001, 0);
INSERT INTO Foobar VALUES (200, 2002, 0);
INSERT INTO Foobar VALUES (300, 2001, 0);
INSERT INTO Foobar VALUES (300, 2002, 0);
INSERT INTO Foobar VALUES (300, 2003, 0);
INSERT INTO Foobar VALUES (300, 2004, 0);
INSERT INTO Foobar VALUES (300, 2005, 0);
SELECT X.tid, MIN(X.yr), X.end_yr, (X.end_yr - MIN(X.yr)) AS duration
FROM
(SELECT F1.tid, F1.yr, MAX(F2.yr) AS end_yr
FROM Foobar AS F1, Foobar AS F2
WHERE F1.yr < F2.yr
AND F1.tid = F2.tid
AND 0
= ALL (SELECT value
FROM Foobar AS F3
WHERE F1.tid = F3.tid
AND F3.yr BETWEEN F1.yr AND F2.yr)
AND (F2.yr - F1.yr +1)
= (SELECT COUNT(*)
FROM Foobar AS F4
WHERE F1.tid = F4.tid
AND F4.yr BETWEEN F1.yr AND F2.yr)
GROUP BY F1.tid, F1.yr) AS X(tid, yr, end_yr)
GROUP BY X.tid, X.end_yr
HAVING (X.end_yr - MIN(X.yr)) > 1;|||A thoughful solution that leaves one to ponder the rationality
of the methodology and of the environment that spawned it :)
"Trey Walpole" <treypole@.newsgroups.nospam> wrote in message
news:e5k25kf9FHA.2844@.TK2MSFTNGP10.phx.gbl...
> select tid, min(yr) as start, maxyr as [end], count(*) as yrs
> from (
> select *,
> isnull(
> (select max(yr)
> from #test t2
> where t2.tid=t1.tid
> and t2.yr>=t1.yr
> and t2.value=0
> and t2.yr<=(select min(yr) from #test where tid=t2.tid and
> yr>=t1.yr and value=1)
> ),
> (select max(yr)
> from #test t3
> where t3.tid=t1.tid
> )
> ) as maxyr
> from #test t1
> where value=0
> ) t4
> group by tid, maxyr
> having count(*)>=4
> order by tid, start
>
> carmaboy@.gmail.com wrote:|||this problem would be solved like that? ;)
CREATE TABLE Puzzles..YearVals(
tid INTEGER NOT NULL,
yr INTEGER NOT NULL,
value INTEGER NOT NULL,
PRIMARY KEY (tid, yr));
INSERT INTO Puzzles..YearVals VALUES (100, 1998, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 1999, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2000, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2001, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2002, 1);
INSERT INTO Puzzles..YearVals VALUES (100, 2003, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2004, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2005, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2006, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2007, 0);
INSERT INTO Puzzles..YearVals VALUES (100, 2008, 1);
INSERT INTO Puzzles..YearVals VALUES (200, 1999, 0);
--INSERT INTO Puzzles..YearVals VALUES (200, 2000, 1);
INSERT INTO Puzzles..YearVals VALUES (200, 2001, 0);
INSERT INTO Puzzles..YearVals VALUES (200, 2002, 0);
INSERT INTO Puzzles..YearVals VALUES (300, 2001, 0);
INSERT INTO Puzzles..YearVals VALUES (300, 2002, 0);
INSERT INTO Puzzles..YearVals VALUES (300, 2003, 0);
INSERT INTO Puzzles..YearVals VALUES (300, 2004, 0);
INSERT INTO Puzzles..YearVals VALUES (300, 2005, 0);
SELECT tid, MIN(yr) AS beg_yr, MAX(yr) AS fin_yr, COUNT(*) AS tally
FROM
(SELECT B.tid,B.yr,B.yr-COUNT(*)
FROM Puzzles..YearVals AS B
JOIN Puzzles..YearVals AS L
ON L.tid = B.tid AND L.yr <= B.yr AND L.value=B.value
WHERE B.value = 0
GROUP BY B.tid,B.yr) AS G(tid,yr,grp)
GROUP BY grp,tid
HAVING COUNT(*) >= 4|||rrr... i did not remove a reference on my local db
sorry|||Thanks to everyone for taking the time to help. I should have all I
need to get my results.