Counting missing rows in a join?

Question

Darn, I should be able to figure this out,
but have scratched my head for hours on this,
wondering if someone can come up with a
sol'n:
Given two example tables:
Cust Order
cust_no ord_no
cust_no
etc.
I want to be able to list ALL customers,
and the associated number of orders INCLUDING any cases where ZERO orders are registered for any customer.
Cust #Orders
A 13
B 0
C 4
etc
The first part is easy:
SELECT cust_id, count(*) from order
group by cust_id, order by cust_id;
and I can set ZERO values for the "missing"
customers with:
SELECT c.cust_id, 0 from cust c
where not exists
(select * from order
where order.cust_id = c.cust_id);
BUT:
I can't UNION (and then sort) these because
of the stupid GROUP/ORDER clause on the
first query.
Anyone have any suggestion how to get around this?!?

Answer 1

good answers, but as regards group by doing ordering, this is not "correct". It is true that group by needs to order, but you should not rely on this to give you an "ordered" list. You can confirm this caution as it is in the description of the group by clause. If you want to know that your code orders correctly, you must use an order by. If oracle decides he doesn't need the order by because of the grouping, he will exclude it:
create table a (a number);
explain plan for
select a,count(*)
from a
group by a
order by a
SELECT STATEMENT
SORT GROUP BY
TABLE ACCESS FULL KM21378.A
Good luck, Kevin