Question

I have a query that identifies gaps and overlaps of date ranges in sql server 2008 r2. Each unique data set has 12 records. What I would like to do is to adjust or add to the code that identifies the gaps and overlaps and update the records to be sequential.

--gaps and overlaps tbl_volumes
with s as
(
select esiid,Read_Start,Read_End ,row_number() over(partition by esiid order by Read_Start) rn
from tbl_Volumes
where Status=0
group by esiid,Read_Start,Read_End)
select a.esiid, a.Read_Start, a.Read_End, b.Read_Start as nextstartdate,datediff(d,a.Read_End, b.Read_Start) as gap
into #go
from s a
join s b on b.esiid = a.esiid and b.rn = a.rn + 1
where datediff(d, a.Read_End, b.Read_Start) not in (0,1)
order by a.esiid

Here is the bad record set that I would like to see sequential:

e                   Read_Start  Read_End    Source
10032789402145965   2011-01-21  2011-02-22  867_03_1563303
10032789402145965   2011-02-22  2011-03-21  867_03_1665865
10032789402145965   2011-03-26  2011-04-20  867_03_1782993

Answer 1

Well, you could just assign a new Read_end to each record based on the next value. The calculation for the new start can be done like this:

select t.*,
       (select top 1 Read_Start
        from t t2
        where t2.e = t.e and t2.Read_Start > t.Read_Start
        order by t2.Read_Start
       ) as New_Read_End
from t

Do you actually want to update the value or just see what it should be?