Perform SUM, and CASE WHEN - sql

Question

I'm using SQL Server 2008.
I get this error:
is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.
I would like to:
Select all columns from table b
When there is the same numberID I would like to perform SUM and return one row with result
Also, if there is just one numberID I would like to return result
I would like to create one more column Case with the following output “Percentage is…” (value from column Percentage)
Code so far:
select distinct
b.*,
case
when b.NumberID = b.NumberID then SUM(d.Percentage)
end as Percentage,
NULL as LName,
NULL as FName,
'Percentage is' + convert(VARCHAR(20),
CASE WHEN b. NumberID = b.NumberID
THEN SUM(d.Percentage) END) as Case
from
Accounting b
join (
select b.*
from Accounting b
join (
select number
, MAX(id) id
from Accounting
where Date < '2012-12-01'
group by number
) a
on b.number = a.number
and b.Id = a.id
) b2
on b.number = b2.number
where b.Date > '2012-12-01'
group by b.NumberID

Answer 1

Use distinct in the SELECT under join as well.
//etc.
join
(
select DISTINCT b.* from
Accounting b
//etc.

Answer 2

You case syntax is incorrect. Check the manual.
It should go like this:
CASE case_value
WHEN when_value
so in your case (jeje)
select distinct b.*,
case b.NumberID when b.NumberID then ...
which is strange indeed. b.NumberID is always b.NumberID :o

Answer 3

SELECT b.NumberID, SUM(b.Percentage) AS Percentage,
(
SELECT MAX(s.id)
FROM Accounting s
WHERE s.Date < '2012-12-01' AND s.number = b.number
) AS MaxID
FROM Accounting b
WHERE b.Date > '2012-12-01'
GROUP BY b.NumberID

Answer 4

You need all of the columns in your select statement included in the Group By clause. Let's say I have a table of ice cream favors and people that ordered them. I have data that says
Joe ordered chocolate
Katie ordered Strawberry
Mary ordered chocolate
If I tell sql to group by the ice cream flavor, I am telling it that I only want one row to be returned for that flavor. By including the customer's name in my select, I am asking sql who ordered each of the flavors. How can I show (in one row) who ordered chocolate? In order to ask sql for the information that I want, I need a row for each customer and each ice cream flavor (Maybe Mary ordered strawberry yesterday). So, I need to group by both customer and ice cream flavor.
Agregate functions don't need to be in the group by clause because you are asking sql a specific question regarding the data. Maybe I want to know the person whose name comes first alphabetically. Then, I could select the max(customer) and group by just the ice cream flavor.
Select b.id, b.number, b.firmname, b.date, sum(...)
From
Group by b.id, b.number, b.firmname, b.date
Notes:
You can still use the b.* syntax, as long as you name all the columns
in your group by
The distinct keyword is no longer necessary in your select, since
group by does essentially the same thing