Query: Assign Cycle Id

Question

Hello every one,
I need help in the query
WITH T AS
     (SELECT 2006 ID,
             1 st,
             11 num
        FROM DUAL
      UNION ALL
      SELECT 2006,
             2,
             12
        FROM DUAL
      UNION ALL
      SELECT 2006,
             5,
             13
        FROM DUAL
      UNION ALL
      SELECT 2006,
             1,
             14
        FROM DUAL
      UNION ALL
      SELECT 2006,
             2,
             15
        FROM DUAL
      UNION ALL
      SELECT 2008,
             1,
             11
        FROM DUAL
      UNION ALL
      SELECT 2007,
             5,
             11
        FROM DUAL
      UNION ALL
      SELECT 2007,
             2,
             12
        FROM DUAL
      UNION ALL
      SELECT 2007,
             3,
             13
        FROM DUAL
      UNION ALL
      SELECT 2007,
             5,
             14
        FROM DUAL
      UNION ALL
      SELECT 2007,
             2,
             15
        FROM DUAL)
SELECT   *
    FROM T
ORDER BY 1,
         3
ID       ST     NUM
2006     1     11
2006     2     12
2006     5     13
2006     1     14
2006     2     15
2007     5     11
2007     2     12
2007     3     13
2007     5     14
2007     2     15
2008     1     11
Assign cycle id group by id order by num, assing new cycle per group after every occurance of 5 in st column. Measn rows before 5 and should get same cyple id
the data should look like this.
ID       ST     NUM     Cycle id
2006     1     11     1   ^
2006     2     12     1   ^
2006     5     13     1   ^ <<- cycle 1
2006     1     14     2  
2006     2     15     2     <<- cycle 2
2007     5     11     1
2007     2     12     2
2007     3     13     2
2007     5     14     2
2007     2     15     3
2008     1     11     1
         Thank you

Answer 1

Hi,
With 10g
Regards Salim.
SELECT  id,st,num,  cycle_id
    FROM T
model
partition by (id)
dimension by (row_number()over(partition by id order by num)rn)
measures( st,num, 1 cycle_id)
rules
(cycle_id[rn] order by rn asc= case when st[cv()-1] is null then 1
                                    when st[cv()-1]=5 then cycle_id[cv()-1]+1 
                                    else cycle_id[cv()-1]
                                    end )
order by id,num;
WITH T AS
     (SELECT 2006 ID,
             1 st,
             11 num
        FROM DUAL
      UNION ALL
      SELECT 2006,
             2,
             12
        FROM DUAL
      UNION ALL
      SELECT 2006,
             5,
             13
        FROM DUAL
      UNION ALL
      SELECT 2006,
             1,
             14
        FROM DUAL
      UNION ALL
      SELECT 2006,
             2,
             15
        FROM DUAL
      UNION ALL
      SELECT 2008,
             1,
             11
        FROM DUAL
      UNION ALL
      SELECT 2007,
             5,
             11
        FROM DUAL
      UNION ALL
      SELECT 2007,
             2,
             12
        FROM DUAL
      UNION ALL
      SELECT 2007,
             3,
             13
        FROM DUAL
      UNION ALL
      SELECT 2007,
             5,
             14
        FROM DUAL
      UNION ALL
      SELECT 2007,
             2,
             15
        FROM DUAL)
SELECT  id,st,num,  cycle_id
    FROM T
model
partition by (id)
dimension by (row_number()over(partition by id order by num)rn)
measures( st,num, 1 cycle_id)
rules
(cycle_id[rn] order by rn asc= case when st[cv()-1] is null then 1
                                    when st[cv()-1]=5 then cycle_id[cv()-1]+1 
                                    else cycle_id[cv()-1]
                                    end )
order by id,num;
        ID         ST        NUM   CYCLE_ID
      2006          1         11          1
      2006          2         12          1
      2006          5         13          1
      2006          1         14          2
      2006          2         15          2
      2007          5         11          1
      2007          2         12          2
      2007          3         13          2
      2007          5         14          2
      2007          2         15          3
      2008          1         11          1
11 rows selected.Edited by: Salim Chelabi on 2009-06-19 11:29