Group data by "week" in 3d Chart

Question

Hi all,
I need to be able to group the results of the below query by week, so it displays better - problem is I have no idea how to do this.
At the moment it is finding all "active" projects in my database, and then displaying the count that are deploying on a given day. What I now need to do is display the total projects deploying in a given "week" over the date range (ie. Week 1 has 23, Week 2 has 13, and so on)
Can anyone advise how I would go about modifiying the query to achieve this in Apex 4.1?
select
'f?p=&APP_ID.:87:&APP_SESSION.::NO:RP:P87_DISPLAY_DATE:' || bus_deployment_launch link,
bus_deployment_launch LABEL,
sum(No_Proj) Value
from
select
  count(a.initiative_name) No_Proj
  ,a.bus_deployment_launch bus_deployment_launch
  ,a.status status
from EWCM_INITIATIVE a
where (a.STREAM_SERVE like '%Assurance%' or a.STREAM_SERVE like '%O2A%')
group by bus_deployment_launch, status
union
select
   count(name) No_Proj
  ,requested_date bus_deployment_launch
  ,status status
from ewcm_gtm_initiatives
where PORTFOLIO != '141'
and (impact_assurance = 'Y' or impact_o2a = 'Y')
group by requested_date, status
union
select
0 as No_Proj, nvl(to_date(:P85_START_DATE,'dd-mon-yyyy'), TRUNC(SYSDATE,'MONTH')) + level-1  bus_deployment_launch,'In progress' status
     from dual
     connect by level <= nvl(to_date(:P85_END_DATE,'dd-mon-yyyy'),(LAST_DAY(TRUNC(SYSDATE)))) - nvl(to_date(:P85_START_DATE,'dd-mon-yyyy'), TRUNC(SYSDATE,'MONTH'))  +1
where bus_deployment_launch between nvl(:P85_START_DATE,(LAST_DAY(ADD_MONTHS(TRUNC(SYSDATE),-1))+1)) and nvl(:P85_END_DATE,(LAST_DAY(TRUNC(SYSDATE))))
   and status in ('In progress','Non-GTM','Ideas Stage','Scope Review','Change Management','Change Mgmt','Pending','Ongoing')
group by bus_deployment_launch
order by label asc

Answer 1

Paul Ferris wrote:
Sorry I mean to show a count of projects in a week (so adding Monday through to Sunday together)
The simple option—providing it matches your definition of "week"—is just to wrap your existing query in another that aggregates by ISO week (the 'IW' format mask). With start date 01-JAN-2014 and end date 31-JAN-2014:
with t as (
      select
          0 no_proj
        ,   nvl(
                to_date(:p85_start_date, 'dd-mon-yyyy')
              , trunc(sysdate, 'MONTH'))
          + level - 1 bus_deployment_launch
        , 'In progress' status
      from
        dual
          connect by level <=   nvl(
                                    to_date(:p85_end_date, 'dd-mon-yyyy')
                                  , last_day(trunc(sysdate)))
                              - nvl(to_date(:p85_start_date, 'dd-mon-yyyy'), trunc(sysdate, 'MONTH')) + 1)
select
    to_char(trunc(bus_deployment_launch, 'iw'), 'DY DD.MM.YYYY') week
  , row_number()
      over (
        order by trunc(bus_deployment_launch, 'iw')) week#
  , status
  , count(*) n
from
    t
group by
    status
  , trunc(bus_deployment_launch, 'iw')
WEEK                         WEEK# STATUS               N
MON 30.12.2013                   1 In progress          5
MON 06.01.2014                   2 In progress          7
MON 13.01.2014                   3 In progress          7
MON 20.01.2014                   4 In progress          7
MON 27.01.2014                   5 In progress          5
You should be able to remove the aggregation at day level from the inner queries.