Using GROUP BY ROLL UP can I sum only certain levels?

Question

I am using a group by rollup in the hope to add a grand total.
Here's an example
with mydata as (
select 'A' part_no, 'A Part' Description, '100' serial, 10 Qty, 45 unit_cost from dual
UNION ALL
select 'A' part_no, 'A Part' Description, '101' serial, 10 Qty, 45 unit_cost from dual
UNION ALL
select 'A' part_no, 'A Part' Description, '102' serial, 5 Qty, 46 unit_cost from dual
UNION ALL
select 'A' part_no, 'A Part' Description, '103' serial, 10 Qty, 45 unit_cost from dual
UNION ALL
select 'B' part_no, 'B Part' Description, '200' serial, 15 Qty, 10 unit_cost from dual
UNION ALL
select 'C' part_no, 'C Part' Description, '100' serial, 1 Qty, 100 unit_cost from dual
UNION ALL
select 'C' part_no, 'C Part' Description, '101' serial, 10 Qty, 95 unit_cost from dual
select
part_no,
description,
sum(qty) total_qty,
sum(qty*unit_cost) total_value
from mydata
group by
rollup(
part_no,
description)However, I only want a GRAND total, not subtotals.
I can use the clause at the end:
having
(grouping(part_no)=0 and grouping(description)=0)
or
(grouping(part_no)=1 and grouping(description)=1)But is there a simpler way?
I was using this kind of method in the past:
with mydata as (
select 'A' part_no, 'A Part' Description, '100' serial, 10 Qty, 45 unit_cost from dual
UNION ALL
select 'A' part_no, 'A Part' Description, '101' serial, 10 Qty, 45 unit_cost from dual
UNION ALL
select 'A' part_no, 'A Part' Description, '102' serial, 5 Qty, 46 unit_cost from dual
UNION ALL
select 'A' part_no, 'A Part' Description, '103' serial, 10 Qty, 45 unit_cost from dual
UNION ALL
select 'B' part_no, 'B Part' Description, '200' serial, 15 Qty, 10 unit_cost from dual
UNION ALL
select 'C' part_no, 'C Part' Description, '100' serial, 1 Qty, 100 unit_cost from dual
UNION ALL
select 'C' part_no, 'C Part' Description, '101' serial, 10 Qty, 95 unit_cost from dual
select
part_no,
description,
sum(qty) total_qty,
sum(qty*unit_cost) total_value
from
mydata
group by
part_no,
description
UNION ALL
select
null,
null,
sum(qty) total_qty,
sum(qty*unit_cost) total_value
from
mydataWhich is inefficient, especially over a lot of data. But isnt aggregating over never to be used levels inefficient too?
Thanks

Answer 1

Using Composite column:
SQL> with mydata as (
  2    select 'A' part_no, 'A Part' Description, '100' serial, 10 Qty, 45 unit_cost from dual
  3    UNION ALL
  4    select 'A' part_no, 'A Part' Description, '101' serial, 10 Qty, 45 unit_cost from dual
  5    UNION ALL
  6    select 'A' part_no, 'A Part' Description, '102' serial, 5 Qty, 46 unit_cost from dual
  7    UNION ALL
  8    select 'A' part_no, 'A Part' Description, '103' serial, 10 Qty, 45 unit_cost from dual
  9    UNION ALL
10    select 'B' part_no, 'B Part' Description, '200' serial, 15 Qty, 10 unit_cost from dual
11    UNION ALL
12    select 'C' part_no, 'C Part' Description, '100' serial, 1 Qty, 100 unit_cost from dual
13    UNION ALL
14    select 'C' part_no, 'C Part' Description, '101' serial, 10 Qty, 95 unit_cost from dual)
15  select part_no,description,sum(qty) total_qty,sum(qty*unit_cost) total_value
16  from mydata
17  group by rollup((part_no,description))
18  ;
P DESCRI  TOTAL_QTY TOTAL_VALUE
A A Part         35        1580
B B Part         15         150
C C Part         11        1050
                 61        2780Regards,
Dima