Sort Records comparing sums across multiple tables - sql

Question

In SQL Server, I am wanting to bring back all jobs where
(
SUM(Order.InvoicesReceived) > (SUM(Estimate.GrossValue) + SUM (AdditionalEstimate.GrossValue))
) OR (
SUM(Order.ContractGiven) > (SUM(Estimate.GrossValue) + SUM (AdditionalEstimate.GrossValue))
)
CREATE TABLE Job (id INT, userid INT)
INSERT INTO Job ( id ,userid)VALUES ( 1,1)
INSERT INTO Job ( id ,userid)VALUES ( 2,1)
INSERT INTO Job ( id ,userid)VALUES ( 3,2)
INSERT INTO Job ( id ,userid)VALUES ( 4,2)
INSERT INTO Job ( id ,userid)VALUES ( 5,1)
CREATE TABLE [User] (id INT, UserName NVARCHAR (30))
INSERT INTO [User] ( id ,UserName)VALUES ( 1,'Richard')
INSERT INTO [User] ( id ,UserName)VALUES ( 2,'Jane')
CREATE Table Estimate (id INT, [job] INT, [GrossValue] DECIMAL (18,2))
INSERT INTO Estimate ( id ,[job], GrossValue)VALUES ( 1,3, 100)
INSERT INTO Estimate ( id ,[job], GrossValue)VALUES ( 2,4, 100)
INSERT INTO Estimate ( id ,[job], GrossValue)VALUES ( 3,5, 200)
INSERT INTO Estimate ( id ,[job], GrossValue)VALUES ( 4,5, 200)
CREATE Table AdditionalEstimate (id INT, [job] INT, [GrossValue] DECIMAL (18,2))
INSERT INTO AdditionalEstimate ( id ,[job], GrossValue)VALUES ( 1,1, 100)
INSERT INTO AdditionalEstimate ( id ,[job], GrossValue)VALUES ( 2,2, 100)
INSERT INTO AdditionalEstimate ( id ,[job], GrossValue)VALUES ( 3,5, 100)
INSERT INTO AdditionalEstimate ( id ,[job], GrossValue)VALUES ( 4,5, 100)
CREATE Table [Order] (id INT, [job] INT, ContractGiven DECIMAL (18,2), InvoicesReceived DECIMAL (18,2))
INSERT INTO [Order] ( id ,[job], ContractGiven,InvoicesReceived)VALUES ( 1,1, 50, 0)
INSERT INTO [Order] ( id ,[job], ContractGiven,InvoicesReceived)VALUES ( 2,2, 150, 0)
INSERT INTO [Order] ( id ,[job], ContractGiven,InvoicesReceived)VALUES ( 3,3, 50, 0)
INSERT INTO [Order] ( id ,[job], ContractGiven,InvoicesReceived)VALUES ( 4,4, 150, 0)
INSERT INTO [Order] ( id ,[job], ContractGiven,InvoicesReceived)VALUES ( 5,5, 400, 0)
INSERT INTO [Order] ( id ,[job], ContractGiven,InvoicesReceived)VALUES ( 6,5, 100, 0)
To make it easy to see what results I should get I added the following table and updated the inserts into tables.
JOB Estimate AddEstimate Order
1 Null 100 50
2 Null 100 150
3 100 NULL 50
4 100 NULL 150
5 200 100 500
5 200 100 NA
InvoicedRecieved is ignored for simplicity. Jobs 2,4 Should be returned.
Richard 1, Jane 1
I also need another statement bring back the number of jobs per user that met the above criteria.

Answer 1

Gordon was correct about the need to sum before comparing, due to the Cartesian product. However as his query isn't producing the correct result here is the query which does.
select J.id [Job]
, coalesce(O.InvoicesReceived,0) InvoicesReceived
, coalesce(O.ContractGiven,0) ContractGiven
, coalesce(E.GrossValue,0) + coalesce(A.GrossValue,0) TotalQuoteCost
from Job J
left join (select Job, SUM(O.InvoicesReceived) InvoicesReceived, SUM(O.ContractGiven) ContractGiven from [Order] O group by Job) O on O.Job = J.id
left join (select Job, SUM(E.GrossValue) GrossValue from Estimate E group by Job) E on E.Job = J.id
left join (select Job, SUM(A.GrossValue) GrossValue from AdditionalEstimate A group by Job) A on A.Job = J.id
where (
coalesce(O.InvoicesReceived,0) > (coalesce(E.GrossValue,0) + coalesce(A.GrossValue,0))
) OR (
coalesce(O.ContractGiven,0) > (coalesce(E.GrossValue,0) + coalesce(A.GrossValue,0))
)
select [User Name], count(*)
from (
select U.UserName [User Name], J.id
, coalesce(O.InvoicesReceived,0) InvoicesReceived
, coalesce(O.ContractGiven,0) ContractGiven
, coalesce(E.GrossValue,0) + coalesce(A.GrossValue,0) TotalQuoteCost
from Job J
inner join [User] U on U.id = J.UserId
left join (select Job, SUM(O.InvoicesReceived) InvoicesReceived, SUM(O.ContractGiven) ContractGiven from [Order] O group by Job) O on O.Job = J.id
left join (select Job, SUM(E.GrossValue) GrossValue from Estimate E group by Job) E on E.Job = J.id
left join (select Job, SUM(A.GrossValue) GrossValue from AdditionalEstimate A group by Job) A on A.Job = J.id
where (
coalesce(O.InvoicesReceived,0) > (coalesce(E.GrossValue,0) + coalesce(A.GrossValue,0))
) OR (
coalesce(O.ContractGiven,0) > (coalesce(E.GrossValue,0) + coalesce(A.GrossValue,0))
)
) x
group by [User Name]
DB Fiddle

Answer 2

You need to aggregate before you join:
select U.UserName, count(*)
from [User] U join
Job j
on u.id = j.userid left join
(select o.job, sum(O.InvoicesReceived) as InvoicesReceived,
sum(o.ContractGiven) as ContractGiven
from [Order] O
group by o.job
) o
on o.Job = j.id left join
(select e.job, sum(e.GrossValue) as GrossValue
from estimate e
group by e.job
) e
on e.Job = j.id left join
(select ae.job, sum(ae.GrossValue) as GrossValue
from AdditionalEstimate ae
group by ae.job
) ae
on ae.Job = j.id
group by U.UserName
having coalesce(sum(o.InvoicesReceived), 0) > coalesce(sum(e.GrossValue), 0) + coalesce(sum(ae.GrossValue), 0) or
coalesce(sum(o.ContractGiven), 0) > coalesce(sum(e.GrossValue), 0) + coalesce(sum(ae.GrossValue), 0);
Otherwise, the joins will produce a Cartesian product and the sum()s will be off.
Here is a db<>fiddle.
If you want the jobs that meet these conditions:
select U.UserName, j.id
from [User] U join
Job j
on u.id = j.userid left join
(select o.job, sum(O.InvoicesReceived) as InvoicesReceived,
sum(o.ContractGiven) as ContractGiven
from [Order] O
group by o.job
) o
on o.Job = j.id left join
(select e.job, sum(e.GrossValue) as GrossValue
from estimate e
group by e.job
) e
on e.Job = j.id left join
(select ae.job, sum(ae.GrossValue) as GrossValue
from AdditionalEstimate ae
group by ae.job
) ae
on ae.Job = j.id
where coalesce(o.InvoicesReceived, 0) > coalesce(e.GrossValue, 0) + coalesce(ae.GrossValue, 0) or
coalesce(o.ContractGiven, 0) > coalesce(e.GrossValue, 0) + coalesce(ae.GrossValue, 0);
Which has this db<>fiddle.