Having Trouble Summing distinct values - sql

Question

I have a fairly poorly designed DB that I'm trying to pull reports from. I'm attempting to sum the value on the column GuestCount, however with the structure of the joins, i'm getting a cartesian situation that's making the sum inaccurate. I can't use Sum(Distinct) because I'm not trying to sum the distinct values in GuestCount, but rather the sum of distinct rows.
Here's the SQL to set up the Tables:
CREATE TABLE TesttblTransactions (
ID int,
[sysdate] date,
TxnHour tinyint,
Facility nvarchar(50),
TableID int,
[Check] int,
Item int,
Parent int
)
Create Table TesttblTablesGuests (
ID int,
Facility nvarchar(50),
TableID int,
GuestCount tinyint,
TableDate Date
)
Create Table TesttblFacilities (
ID int,
ClientKey nvarchar(50),
Brand nvarchar(50),
OrgFacilityID nvarchar(50),
UnitID smallint
)
INSERT INTO testtbltransactions (
ID,
[Sysdate],
TxnHour,
Facility,
TableID,
[Check],
Item,
Parent
)
VALUES
(
1,
'20221201',
7,
'JOES',
1001,
12345,
8898989,
0
),
(
2,
'20221201',
7,
'JOES',
1001,
12345,
8776767,
1
),
(
3,
'20221201',
7,
'JOES',
1001,
12345,
856643,
0
),
(
4,
'20221201',
7,
'THE DIVE',
1001,
67890,
662342,
0
),
(
5,
'20221201',
7,
'THE DIVE',
1001,
67890,
244234,
0
),
(
6,
'20221201',
7,
'JOES',
1002,
12344,
873323,
0
);
INSERT INTO testtblTablesGuests (
ID,
Facility,
TableID,
GuestCount,
TableDate
)
VALUES
(
1,
'JOES',
1001,
4,
'20221201'
),
(
2,
'THE DIVE',
1001,
1,
'20221201'
),
(
3,
'JOES',
1002,
1,
'20221201'
);
INSERT INTO testtblFacilities (
ID,
ClientKey,
Brand,
OrgFacilityID,
UnitID
)
VALUES
(
1,
'JOES',
'Joes Hospitality Group LLC',
'Joes Bar',
987
),
(
2,
'THE DIVE',
'The Dive Restaurant Group',
'The Dive',
565
);
--Here's the SQL that I need for reporting but can't seem to get working:
Declare #StartDate as Date = '12-1-2022'
Declare #EndDate as Date = '12-1-2022'
--The query we want to work
SELECT
TesttblFacilities.ClientKey,
TesttblFacilities.Brand,
format(testtbltransactions.sysdate,'yyyy-MM-dd') AS [Date],
'H' AS Freqency,
Testtbltransactions.[TxnHour] AS [Hour],
TesttblFacilities.UnitID AS [UnitID],
'Dine In Guest Count' as Metric,
Sum(TesttblTablesGuests.GuestCount) AS [Value]
FROM ((Testtbltransactions
JOIN Testtbltablesguests ON (Testtbltablesguests.TableDate = Testtbltransactions.sysdate) AND (Testtbltransactions.FACILITY = Testtbltablesguests.facility) AND (Testtbltransactions.tableid = Testtbltablesguests.tableid))
JOIN TesttblFacilities ON Testtbltransactions.FACILITY = TesttblFacilities.ClientKey)
Where (((Testtbltransactions.parent)=0))
and Testtbltransactions.sysdate >= #StartDate
and Testtbltransactions.sysdate <= #EndDate
GROUP BY TesttblFacilities.ClientKey, Testtblfacilities.UnitID,TesttblFacilities.Brand, Testtbltransactions.facility, Testtbltransactions.sysdate, Testtbltransactions.TxnHour`
I'm getting 9 and 2, instead of 5 and 1.
In the comments NBK suggested doing a subquery - and it took me a while but I think i found something that works.. . .
Declare #StartDate as Date = '12-1-2022'
Declare #EndDate as Date = '12-1-2022'
Select
t1.txnhour,
t1.facility,
SUM(t1.guestcount) from
(
Select Distinct
TesttblTransactions.TableID as TableID,
testtbltransactions.[txnHour] as txnhour,
testtbltransactions.Facility as Facility,
testtbltablesguests.GuestCount as guestcount,
testtbltransactions.Parent as parent
From TesttblTransactions
Join TesttblTablesGuests on TesttblTablesGuests.TableID = TesttblTransactions.TableID and testtbltablesguests.Facility = TesttblTransactions.Facility
Where (((Testtbltransactions.parent)=0))
and Testtbltransactions.sysdate >= #StartDate
and Testtbltransactions.sysdate <= #EndDate
) T1
Group by t1.Facility, t1.txnhour, t1.Facility
I'm going to continue to refine this, but I think I should be able to move forward with this.