Difficulty working with overlapping date ranges and sums

Question

I have data like the following and need help querying it for a specific result:
startdate             enddate               place     class       number
03-30-2006          04-18-2006          A             22           2000
04-01-2006          04-30-2006          A             22           6000
03-30-2006          04-15-2006          B             11           1700
03-15-2006          04-12-2006          B             11           2900
I'd like to return the sum of the averages for any given day, by place, and by class, like this:
day                    place     class       average
03-15-2006          B             11           100
03-16-2006          B             11           100
03-30-2006          A             22           100
03-30-2006          B             11           200
03-31-2006          A             22           100
03-31-2006          B             11           200
04-1-2006            A             22           300
04-01-2006          B             11           200
04-30-2006          A             11           200
I'm stuck, any help would be GREATLY appreciated.

Answer 1

307b218a-cebe-46f1-87b6-946b4076ce79 wrote:
Here is what I said in my original post:
I'd like to return the sum of the averages for any given day, by place, and by class.....Then gave an example of output.
I think that sentence describes the relations perfectly succinctly.
How can it, when there's no SUM involved
307b218a-cebe-46f1-87b6-946b4076ce79 wrote:
For the date of 3-30-2006, the sum of the averages for region A in class 22 is 100, since from 3-30-2006 to 4-18-2006 is 20 days, and 2000/20 = 100, and there are no other class 22's in place A in that date range. Same goes for 3-31-20006.
However for 4-1-2006 the sum of the averages is 300, since I have one record who's average is 100 (2000/20), and another (6000/30 = 200) and 200 + 100 = 300.
For 4/30, I have one average in that range for place A class 22, and its average for that period is 200.
Does that help?
Much better, thank-you ..
try this:
WITH data AS (
      SELECT TO_DATE('03-30-2006','mm-dd-yyyy') startdate, TO_DATE('04-18-2006','mm-dd-yyyy') enddate, 'A' place, 22 class, 2000 cnumber FROM DUAL UNION ALL
      SELECT TO_DATE('04-01-2006','mm-dd-yyyy') startdate, TO_DATE('04-30-2006','mm-dd-yyyy') enddate, 'A' place, 22 class, 6000 cnumber FROM DUAL UNION ALL
      SELECT TO_DATE('03-30-2006','mm-dd-yyyy') startdate, TO_DATE('04-15-2006','mm-dd-yyyy') enddate, 'B' place, 11 class, 1700 cnumber FROM DUAL UNION ALL
      SELECT TO_DATE('03-15-2006','mm-dd-yyyy') startdate, TO_DATE('04-12-2006','mm-dd-yyyy') enddate, 'B' place, 11 class, 2900 cnumber FROM DUAL ),
   data2 AS ( SELECT startdate, enddate, enddate-startdate+1 numdays, place, class, cnumber
               FROM data d
SELECT startdate+x.lvl day, place, class, cnumber/numdays avg
  FROM data2 d,
      ( SELECT level lvl FROM dual CONNECT BY level <= 10000 ) x
WHERE x.lvl <= numdays
ORDER BY d.startdate, x.lvl
It helps to be specific and detailed