Question

I’m getting an unexpected result for my query in group function. I’m using the following 3 tables:

1. sale with columns AccountId, NetAmount, quantity, date
2. Purchase columns AccountId, NetAmount, quantity, date
3. Account columns AccountId, AccountName

I made a stored procedure that takes two inputs: Date1 and Date2.

I need to calculate the following:

• Account.AccountName
• sum of NetAmount of Purchase
• sum of NetAmount of Sale with Date before Date
• sum of NetAmount of Purchase – sum of NetAmount of Sale for Date between Date1 and Date2
• sum of NetAmount of Sale and Purchase for Date between Date1 and Date2

I’m currently doing this:

SELECT a.SecurityName,
       Sum( d.NetAmount) - Sum(e.NetAmount)As 'Opening Amount',
       Sum( d.Quantity) - Sum(e.Quantity) As 'Opening Number',
       Sum( d.NetAmount) / Sum( d.Quantity)As 'Opening Rate', 
       Sum( s.Quantity) As 'Number', 
       Sum( s.NetAmount) / Sum( s.Quantity) As 'Rate', 
       Sum( s.NetAmount) As 'Amount',
       Sum( p.Quantity) As 'Number',
       Sum( p.NetAmount) / Sum( p.Quantity) As 'Rate',
       Sum(  p.NetAmount) AS 'Amount', 
       IsNull(Sum( d.Quantity), 0) + (Sum( p.Quantity)) - IsNull((Sum( s.Quantity)), 0) As 'Closing Number',
       IsNull(Sum( d.NetAmount),0)+(Sum( p.NetAmount)) -IsNull((Sum( s.NetAmount)),0) As 'Closing Amount',
       IsNull(Sum( d.Rate),0)+(Sum( p.Rate))-IsNull((Sum( s.Rate)),0) As 'Closing Rate'
  FROM Sale s 
left Join SecurityAccount a ON s.SecurityAccountId = a.SecurityAccountId 
Right JOIN Purchase p ON a.SecurityAccountId = p.SecurityAccountId 
Left JOin Purchase d On a.SecurityAccountId=d.SecurityAccountId
                    And d.Date < @PeriodStart
Left Join Sale e On a.SecurityAccountId=e.SecurityAccountId
                And e.Date < @PeriodStart
Group by a.SecurityName

End

But I’m getting values 3 times greater than expected.

Can anyone tell me what should i do?

Answer 1

You are joining tables 4 times by the same field SecurityAccountId. Each join will result multiplying of result rows. The only way I see is to create 4 subqueries with grouping and then using those results in main query. This should work, if I have no mistakes 🙂

SELECT a.SecurityName,
       (d.SumNetAmount - e.SumNetAmount) AS 'Opening Amount',
       (d.SumQuantity - e.SumQuantity) AS 'Opening Number',
       (d.SumNetAmount) / d.SumQuantity) AS 'Opening Rate', 
       s.SumQuantity AS 'Number', 
       (s.SumNetAmount / s.SumQuantity) AS 'Rate', 
       s.SumNetAmount AS 'Amount',
       p.SumQuantity AS 'Number',
       (p.SumNetAmount / p.SumQuantity) AS 'Rate',
       p.SumNetAmount AS 'Amount', 
       (ISNULL(d.SumQuantity, 0) + p.SumQuantity - ISNULL(s.SumQuantity, 0)) AS 'Closing Number',
       (ISNULL(d.SumNetAmount,0) + p.SumNetAmount - ISNULL(s.SumNetAmount,0)) AS 'Closing Amount',
       (ISNULL(d.SumRate,0) + p.SumRate - ISNULL(s.SumRate,0)) As 'Closing Rate'
FROM SecurityAccount a
LEFT JOIN (SELECT SecurityAccountId, 
       SUM(Quantity) AS 'SumQuantity', 
       SUM(NetAmount) AS 'SumNetAmount',
       SUM(Rate) AS 'SumRate'
FROM Sale) AS s ON a.SecurityAccountId = s.SecurityAccountId
LEFT JOIN (SELECT SecurityAccountId, 
       SUM(Quantity) AS 'SumQuantity', 
       SUM(NetAmount) AS 'SumNetAmount',
       SUM(Rate) AS 'SumRate'
FROM Sale WHERE Date < @PeriodStart) AS e ON a.SecurityAccountId = e.SecurityAccountId
LEFT JOIN (SELECT SecurityAccountId, 
       SUM(Quantity) AS 'SumQuantity', 
       SUM(NetAmount) AS 'SumNetAmount',
       SUM(Rate) AS 'SumRate'
FROM Purchase) AS p ON a.SecurityAccountId = p.SecurityAccountId
LEFT JOIN (SELECT SecurityAccountId, 
       SUM(Quantity) AS 'SumQuantity', 
       SUM(NetAmount) AS 'SumNetAmount',
       SUM(Rate) AS 'SumRate'
FROM Purchase WHERE Date < @PeriodStart) AS d ON a.SecurityAccountId = d.SecurityAccountId