RE: Analytic troubles

From: Yavor Ivanov <Yavor_Ivanov_at_stemo.bg>
Date: Mon, 10 Nov 2008 13:25:41 +0200
Message-ID: <BD17E2E69E17C64A9684C940EB580E0301102BC74E22@stemodc1.stemo.local>

        Thank you wall. I should be ashamed, I took the wrong path. I was trying to make it with FIRST_VALUE, and the windowing got too messy... While in just e few minutes 7 people gave me a solution with LAG.
        Here is the final solution (including date_to column form the same window):

select name, date_from, lead(prev_date, 1, null) over (order by date_from) date_to   from (select name, date_from,

               lag(name, 1, '*') over(order by date_from) prev_name,
               lag(date_to, 1, null) over(order by date_from) prev_date
          from table1)

 where name != prev_name

Regards,
Yavor Ivanov

-----Original Message-----
From: jo_holvoet_at_amis.com [mailto:jo_holvoet_at_amis.com] Sent: Monday, November 10, 2008 12:12 PM To: Yavor Ivanov
Cc: oracle-l_at_freelists.org; oracle-l-bounce_at_freelists.org Subject: Re: Analytic troubles

select date_from, name from (
select a.date_from, a.name name, lag(a.name, 1) over (order by a.date_from) name1 from table1 a
)
where name <> name1

   or name1 is null

mvg/regards

Jo

             Yavor Ivanov
             <Yavor_Ivanov_at_ste
             mo.bg>                                                     To
             Sent by:                  "oracle-l_at_freelists.org"
             oracle-l-bounce_at_f         <oracle-l_at_freelists.org>
             reelists.org                                               cc

                                                                   Subject
             10/11/2008 10:43          Analytic troubles


             Please respond to
             Yavor_Ivanov_at_stem
                   o.bg






                Hello, Gurus
                I’ve been thrown by the developers in an area, which I do
not visit frequently. Here is the case:
                Let’s say we have a table like this
Date_from                         Name
------------------------------------
01.Nov.2008                      AAA
02.Nov.2008                      AAA
03.Nov.2008                      BBB
04.Nov.2008                      BBB
05.Nov.2008                      BBB
06.Nov.2008                      CCC
07.Nov.2008                      AAA
08.Nov.2008                      AAA

                I need the dates of every change. This is
Date_from         Name
------------------------------------
01.Nov.2008                      AAA
03.Nov.2008                      BBB
06.Nov.2008                      CCC
07.Nov.2008                      AAA

                I can do it with regular SQL, but I make 3 reads of the

table:
select distinct name,

(select min(date_from)

                   from table1 p2
                  where p2.name = p1.name
                    and p2.date_from > (select max(date_from)
                                          from table1 p3
                                         where p3.name != p2.name
                                           and p3.date_from <

p1.date_from)) date_from
  from table1 p1
 I’m trying to implement it with analytic functions (and walk through the table only once), but I cannot define the window based on my knowledge. And window definition is not something very explained in the docs… Can someone help me with this?

(Database is 11g on Windows)

Regards,
Yavor Ivanov
†Ûiÿü0ÁúÞzX¬¶Ê+ƒün– {ú+iÉ^ Received on Mon Nov 10 2008 - 05:25:41 CST