Conditional aggregation

Question

The following represents a daily report reflecting % change in sales where percent diff in sales = ((curr_year_sales - last_year_sales) / last_year_sales) * 100. But district salels for D15-Payne
is only 4.76% although current year sales over last year are huge. store 5601 in that district has unreported
sales and should NOT be included in the district percentage calculation although the other columns should appear as shown. (special reporting rule).
District Sales should be 83.33 % meaning daily sales for last year would effectively be 12000. How do I prevent this. I think im making this hard than whatit is.
DISTRICT                      STORENBR                      DLY_SALES                     DLY_SALES_LY                  DLY_SALES_DIFF                DLY_SALES_PERC                REGION                        STORENAME                     CATEGORYTYPE                
D13-Bowser                    1275                          20000                         11000                         9000                          81.82                         R1-Michaels                   La Vista Ave                  clsStoreRow                 
D13-Bowser                    D13-Bowser                    20000                         11000                         9000                          81.82                         R1-Michaels                   D13-Bowser                    clsDisRow                   
D15-Payne                     2385                          10000                         8000                          2000                          25                            R1-Michaels                   WellGood                      clsStoreRow                 
D15-Payne                     4290                          6000                          3000                          3000                          100                           R1-Michaels                   West Poplar                   clsStoreRow                 
D15-Payne                     5094                          6000                          1000                          5000                          500                           R1-Michaels                   Gtown                         clsStoreRow                 
D15-Payne                     5601                          0                             9000                          -9000                         0                             R1-Michaels                   County Gate                   clsStoreRow                 
D15-Payne                     D15-Payne                     22000                         21000                         1000                          4.76                          R1-Michaels                   D15-Payne                     clsDisRow                   
D29-Bruce                     6705                          20000                         18000                         2000                          11.11                         R1-Michaels                   Levi St                       clsStoreRow                 
D29-Bruce                     D29-Bruce                     20000                         18000                         2000                          11.11                         R1-Michaels                   D29-Bruce                     clsDisRow                   
R1-Michaels                   R1-Michaels                   62000                         50000                         12000                         24                            R1-Michaels                   R1-Michaels                   clsRgnRow                   
Vanco                         Vanco                         62000                         50000                         12000                         24                            Vanco                         Vanco                         clsVancoRow SQL to create data:
with  sales_data_curr_year as
      select 22  storeid ,   153 day_nbr,   10000  dly_sales  from dual union all
      select 23  storeid ,   153 day_nbr,   20000  dly_sales  from dual union all
      select 24  storeid ,   153 day_nbr,    6000  dly_sales  from dual union all
      select 25  storeid ,   153 day_nbr,    6000  dly_sales  from dual union all
      select 27  storeid ,   153 day_nbr,       0  dly_sales  from dual union all
      select 524 storeid ,   153 day_nbr,   20000  dly_sales  from dual
,sales_data_last_year as
      select 22  storeid ,   153 day_nbr,    8000  dly_sales   from dual union all
      select 23  storeid ,   153 day_nbr,   11000  dly_sales   from dual union all
      select 24  storeid ,   153 day_nbr,    3000  dly_sales   from dual union all
      select 25  storeid ,   153 day_nbr,    1000  dly_sales   from dual union all
      select 27  storeid ,   153 day_nbr,    9000  dly_sales   from dual union all
      select 524 storeid ,   153 day_nbr,   18000  dly_sales   from dual
, store_data as
      select 27   storeid,  'County Gate' storename ,    '5601' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy  from dual union all                                                                
      select 25   storeid,  'Gtown' storename ,          '5094' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy  from dual union all                                                              
      select 22   storeid,  'WellGood' storename ,       '2385' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy  from dual union all                                                                  
      select 23   storeid,  'La Vista Ave' storename ,   '1275' storenbr ,     'R1-Michaels' rgn ,   'D13-Bowser' dis , 'BHI' cmpy  from dual union all                                                                 
      select 24   storeid,  'West Poplar' storename ,    '4290' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy  from dual union all                                                                   
      select 524  storeid,  'Levi St' storename ,        '6705' storenbr ,     'R1-Michaels' rgn ,   'D29-Bruce' dis , 'BHI' cmpy  from dual                                                             
--== Decode prevents a field from being null
, rolled_up as
     select    decode(grouping(rgn),       0, rgn,           'Vanco')     as  rgn  
,              decode(grouping(dis),       0, dis,            decode(grouping(rgn),  0,  rgn,        'Vanco'))       as  dis
,              decode(grouping(storenbr),  0, storenbr,       decode(grouping(dis),  0,  dis,         decode(grouping(rgn),   0,   rgn,      'Vanco')))       as storenbr 
,              decode(grouping(storenbr),  0, max(storename), decode(grouping(dis),  0, max(dis),    decode(grouping(rgn),   0,   max(rgn), 'Vanco')))       as storename
--== Use a category type for HTML formatting
,              decode(grouping(storenbr),  0, 'clsStoreRow',  decode(grouping(dis),  0, 'clsDisRow', decode(grouping(rgn),   0,  'clsRgnRow', 'clsVancoRow'))) as categoryType 
,              sum(curr.dly_sales)                     as dly_sales
,              sum(prev.dly_sales)                     as dly_sales_ly
,              sum(curr.dly_sales - prev.dly_sales)    as dly_sales_diff
      from     sales_data_curr_year            curr
      left  outer join sales_data_last_year    prev     on prev.storeid  = curr.storeid
                                                       and prev.day_nbr =  curr.day_nbr
      right outer join store_data              stores   on stores.storeid    = curr.storeid
      group by   rollup(rgn,dis,storenbr)
      order by  rgn, dis, storenbr
select   dis                                             as district
,        storenbr                                        as storenbr
,        dly_sales                                       as dly_sales
,        dly_sales_ly                                    as dly_sales_ly
,        dly_sales_diff                                  as dly_sales_diff
,        case
            when  dly_sales    = 0  then 0
            when  dly_sales_ly = 0  then 0
            else  round((dly_sales_diff / dly_sales_ly) * 100 , 2)
         end                                                  as dly_sales_perc
,        rgn                                             as region
,        storename                                       as storename
,        categoryType                                    as categoryType
from rolled_up;

Answer 1

Thanks. I missed the other fields in the output because they were too far right. Now it makes sense.
Is this the expected result set?
SQL> with  sales_data_curr_year as
  2  (
  3        select 22  storeid ,   153 day_nbr,   10000  dly_sales  from dual union all
  4        select 23  storeid ,   153 day_nbr,   20000  dly_sales  from dual union all
  5        select 24  storeid ,   153 day_nbr,    6000  dly_sales  from dual union all
  6        select 25  storeid ,   153 day_nbr,    6000  dly_sales  from dual union all
  7        select 27  storeid ,   153 day_nbr,    null  dly_sales  from dual union all
  8        select 524 storeid ,   153 day_nbr,   20000  dly_sales  from dual
  9  )
10  ,sales_data_last_year as
11  (
12        select 22  storeid ,   153 day_nbr,    8000  dly_sales   from dual union all
13        select 23  storeid ,   153 day_nbr,   11000  dly_sales   from dual union all
14        select 24  storeid ,   153 day_nbr,    3000  dly_sales   from dual union all
15        select 25  storeid ,   153 day_nbr,    1000  dly_sales   from dual union all
16        select 27  storeid ,   153 day_nbr,    9000  dly_sales   from dual union all
17        select 524 storeid ,   153 day_nbr,   18000  dly_sales   from dual
18  )
19  , store_data as
20  (
21        select 27   storeid,  'County Gate' storename ,    '5601' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy from dual union all
22        select 25   storeid,  'Gtown' storename ,          '5094' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy from dual union all
23        select 22   storeid,  'WellGood' storename ,       '2385' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy from dual union all
24        select 23   storeid,  'La Vista Ave' storename ,   '1275' storenbr ,     'R1-Michaels' rgn ,   'D13-Bowser' dis , 'BHI' cmpy from dual union all
25        select 24   storeid,  'West Poplar' storename ,    '4290' storenbr ,     'R1-Michaels' rgn ,   'D15-Payne' dis , 'BHI' cmpy from dual union all
26        select 524  storeid,  'Levi St' storename ,        '6705' storenbr ,     'R1-Michaels' rgn ,   'D29-Bruce' dis , 'BHI' cmpy from dual
27  )
28  --== Decode prevents a field from being null
29  , rolled_up as
30  (
31       select    decode(grouping(rgn),       0, rgn,           'Vanco')     as  rgn
32  ,              decode(grouping(dis),       0, dis,            decode(grouping(rgn),  0,  rgn,        'Vanco'))       as  dis
33  ,              decode(grouping(storenbr),  0, storenbr,       decode(grouping(dis),  0,  dis,         decode(grouping(rgn),   0, rgn,      'Vanco')))       as storenbr
34  ,              decode(grouping(storenbr),  0, max(storename), decode(grouping(dis),  0, max(dis),    decode(grouping(rgn),   0, max(rgn), 'Vanco')))       as storename
35  --== Use a category type for HTML formatting
36  ,              decode(grouping(storenbr),  0, 'clsStoreRow',  decode(grouping(dis),  0, 'clsDisRow', decode(grouping(rgn),   0,  'clsRgnRow', 'clsVancoRow'))) as categoryType
37  ,              sum(curr.dly_sales)                           as dly_sales
38  ,              sum(nvl2(curr.dly_sales,prev.dly_sales,null)) as dly_sales_ly
39  ,              sum(curr.dly_sales - prev.dly_sales)          as dly_sales_diff
40        from     sales_data_curr_year            curr
41        left  outer join sales_data_last_year    prev     on prev.storeid  = curr.storeid
42                                                         and prev.day_nbr =  curr.day_nbr
43        right outer join store_data              stores   on stores.storeid    = curr.storeid
44        group by   rollup(rgn,dis,storenbr)
45        order by  rgn, dis, storenbr
46  )
47  select   dis                                             as district
48  ,        storenbr                                        as storenbr
49  ,        dly_sales                                       as dly_sales
50  ,        dly_sales_ly                                    as dly_sales_ly
51  ,        dly_sales_diff                                  as dly_sales_diff
52  ,        case
53              when  dly_sales    = 0  then 0
54              when  dly_sales_ly = 0  then 0
55              else  round((dly_sales_diff / dly_sales_ly) * 100 , 2)
56           end                                                  as dly_sales_perc
57  ,        rgn                                             as region
58  ,        storename                                       as storename
59  ,        categoryType                                    as categoryType
60  from rolled_up
61  /
DISTRICT    STORENBR     DLY_SALES DLY_SALES_LY DLY_SALES_DIFF DLY_SALES_PERC REGION      STORENAME    CATEGORYTYP
D13-Bowser  1275             20000        11000           9000          81.82 R1-Michaels La Vista Ave clsStoreRow
D13-Bowser  D13-Bowser       20000        11000           9000          81.82 R1-Michaels D13-Bowser   clsDisRow
D15-Payne   2385             10000         8000           2000             25 R1-Michaels WellGood     clsStoreRow
D15-Payne   4290              6000         3000           3000            100 R1-Michaels West Poplar  clsStoreRow
D15-Payne   5094              6000         1000           5000            500 R1-Michaels Gtown        clsStoreRow
D15-Payne   5601                                                              R1-Michaels County Gate  clsStoreRow
D15-Payne   D15-Payne        22000        12000          10000          83.33 R1-Michaels D15-Payne    clsDisRow
D29-Bruce   6705             20000        18000           2000          11.11 R1-Michaels Levi St      clsStoreRow
D29-Bruce   D29-Bruce        20000        18000           2000          11.11 R1-Michaels D29-Bruce    clsDisRow
R1-Michaels R1-Michaels      62000        41000          21000          51.22 R1-Michaels R1-Michaels  clsRgnRow
Vanco       Vanco            62000        41000          21000          51.22 Vanco       Vanco        clsVancoRow
11 rows selected.I made two adjustments. You said the sales were unreported, but you entered value 0. Unreported in Oracle terminology is NULL, so at line 7 I entered NULL. The second adjustment was to make last years sales evaluate to NULL if this years sales are NULL. You see that happening at line 38.
Regards,
Rob.