SQL Server查询自2017年以来未开具发票的2017年以前的顶级客户端

我正试图写一个查询，列出2017年以前的前200名客户，这些客户自2017年以来没有任何发票金额

我尝试了一些查询的变体，但仍然得到了0个结果，我知道这在默认情况下是不准确的，但我在Excel中获得了正确的信息。我在子查询方面没有太多经验，因此看到了一个解决子查询的机会

SELECT TOP 200
    SUM(OI.OrderInvoiceTotalAmount) AS 'Invoiced',
    O.Deal,
    CASE
       WHEN O.Location LIKE '%Vegas%' THEN 'Las Vegas'
        WHEN O.Location LIKE '%New%' THEN 'New Jersey'
        WHEN O.Location LIKE '%Orange%' THEN 'Los Angeles'
        WHEN O.Location LIKE '%Angeles%' THEN 'Los Angeles'
        WHEN O.Location LIKE '%Angeles%' AND O.OrderType = 'System' THEN 'Las Vegas'
        WHEN O.Location LIKE '%DC%' THEN 'Washington DC'
        WHEN O.Location LIKE '%DC%' AND O.OrderType = 'System' THEN 'New Jersey'
        WHEN O.Location LIKE '%Nashville%' THEN 'Tennessee'
        WHEN O.Location LIKE '%Nashville%' AND O.OrderType = 'System' THEN 'Las Vegas'
        WHEN O.Location LIKE '%Houston%' THEN 'Houston'
        WHEN O.Location LIKE '%Orlando%' THEN 'Orlando'
        WHEN O.Location LIKE 'Penn%' THEN 'Pennsylvania'
        WHEN O.Location='Enterprises' THEN 'Entertainment'
        ELSE O.Location
    END AS 'Location'
FROM 
    dbo.OrderInvoice as OI
INNER JOIN 
    dbo.Orders as O ON O.OrderKey = OI.OrderKey
INNER JOIN 
    dbo.Invoice as I on I.InvoiceKey = OI.InvoiceKey
WHERE 
    O.Status NOT IN ('Snapshot', 'Cancelled', 'Void')
    AND I.InvoiceStatusCode <> 'Estimate'
    AND I.InvoiceType = 'Billing'
    AND Year(I.InvoiceDate) < '2017'
    AND O.Deal NOT LIKE '%4Wall%'
    AND O.Location LIKE '%Nashville%'
GROUP BY
    O.Deal,
    CASE
        WHEN O.Location LIKE '%Vegas%' THEN 'Las Vegas'
        WHEN O.Location LIKE '%New%' THEN 'New Jersey'
        WHEN O.Location LIKE '%Orange%' THEN 'Los Angeles'
        WHEN O.Location LIKE '%Angeles%' THEN 'Los Angeles'
        WHEN O.Location LIKE '%Angeles%' AND O.OrderType = 'System' THEN 'Las Vegas'
        WHEN O.Location LIKE '%DC%' THEN 'Washington DC'
        WHEN O.Location LIKE '%DC%' AND O.OrderType = 'System' THEN 'New Jersey'
        WHEN O.Location LIKE '%Nashville%' THEN 'Tennessee'
        WHEN O.Location LIKE '%Nashville%' AND O.OrderType = 'System' THEN 'Las Vegas'
        WHEN O.Location LIKE '%Houston%' THEN 'Houston'
        WHEN O.Location LIKE '%Orlando%' THEN 'Orlando'
        WHEN O.Location LIKE 'Penn%' THEN 'Pennsylvania'
        WHEN O.Location = 'Enterprises' THEN 'Entertainment'
        ELSE O.Location
    END
HAVING
    (SELECT SUM(OI.OrderInvoiceTotalAmount)
     FROM dbo.OrderInvoice as OI
     INNER JOIN dbo.Invoice as I on OI.InvoiceKey=I.InvoiceKey
     INNER JOIN dbo.Orders as O on OI.OrderKey=O.OrderKey
     WHERE YEAR(I.InvoiceDate) > '2017' 
       AND I.InvoiceStatusCode <> 'Estimate'
       AND I.InvoiceType = 'Billing'
       AND O.Deal NOT LIKE '%4Wall%'
       AND O.Location LIKE '%Nashville%') = 0

如果我需要添加更多信息，请告诉我

---

感谢好心人

在检查缺少的行时，使用“不存在”比实际进行计数和检查是否为0要快。在比较之前，引擎必须对所有行进行计数。我已经评论了我编辑的行。基本上删除了HAVING，并在WHERE上添加了一个NOT EXISTS

---

我假设相关查询的链接是通过交易列。

不相关：当O.Location（如“%Angeles%”时，则为“Los Angeles”；当O.Location（如“%Angeles%”）和O.OrderType='System'时，则“Las Vegas”将永远不会返回“Las Vegas”，因为第一个案例也与O.OrderType='System'匹配。将更具体的案例放在第一位，或将和O.OrderType“System”添加到不太具体的案例中。模拟“%Nashville%”和“%DC%”。开枪！没错，达格。谢谢@StickyBit只要说拥有SumOI.OrderInvoiceTotalAmount=0工具，就可以使用新的物品！谢谢@EZlo，效果很好。哦，是的，谢谢你额外提醒我何时使用不同的别名。

Select TOP 200
Sum(OI.OrderInvoiceTotalAmount) as 'Invoiced',
O.Deal,
Case
    When O.Location Like '%Vegas%' Then 'Las Vegas'
    When O.Location Like '%New%' Then 'New Jersey'
    When O.Location Like '%Orange%' Then 'Los Angeles'
    When O.Location Like '%Angeles%' Then 'Los Angeles'
    When O.Location Like '%Angeles%' AND O.OrderType='System' Then 'Las Vegas'
    When O.Location Like '%DC%' Then 'Washington DC'
    When O.Location Like '%DC%' AND O.OrderType='System' Then 'New Jersey'
    When O.Location Like '%Nashville%' Then 'Tennessee'
    When O.Location Like '%Nashville%' AND O.OrderType='System' Then 'Las Vegas'
    When O.Location Like '%Houston%' Then 'Houston'
    WHen O.Location Like '%Orlando%' Then 'Orlando'
    When O.Location Like 'Penn%' Then 'Pennsylvania'
    When O.Location='Enterprises' Then 'Entertainment'
    Else O.Location
End as 'Location'

From dbo.OrderInvoice as OI
Inner Join dbo.Orders as O on O.OrderKey=OI.OrderKey
Inner Join dbo.Invoice as I on I.InvoiceKey=OI.InvoiceKey

Where O.Status Not In ('Snapshot','Cancelled','Void')
AND I.InvoiceStatusCode<>'Estimate'
AND I.InvoiceType='Billing'
AND Year(I.InvoiceDate)<'2017'
AND O.Deal Not Like '%4Wall%'
AND O.Location Like '%Nashville%'

-- no record from 2017 must exist
AND NOT EXISTS (
    SELECT 
        'no record from 2017'
    From dbo.OrderInvoice as XOI -- Note that we use different alias here so it doesn't get mixed up
    Inner Join dbo.Invoice as XI on XOI.InvoiceKey=XI.InvoiceKey
    Inner Join dbo.Orders as XO on XOI.OrderKey=XO.OrderKey
    Where Year(XI.InvoiceDate)> 2017
        AND XI.InvoiceStatusCode<>'Estimate'
        AND XI.InvoiceType='Billing'
        AND XO.Deal Not Like '%4Wall%'
        AND XO.Location Like '%Nashville%'

        AND O.Deal = XO.Deal -- link between the outmost query and the correlated query
    )

Group By
O.Deal,
Case
    When O.Location Like '%Vegas%' Then 'Las Vegas'
    When O.Location Like '%New%' Then 'New Jersey'
    When O.Location Like '%Orange%' Then 'Los Angeles'
    When O.Location Like '%Angeles%' Then 'Los Angeles'
    When O.Location Like '%Angeles%' AND O.OrderType='System' Then 'Las Vegas'
    When O.Location Like '%DC%' Then 'Washington DC'
    When O.Location Like '%DC%' AND O.OrderType='System' Then 'New Jersey'
    When O.Location Like '%Nashville%' Then 'Tennessee'
    When O.Location Like '%Nashville%' AND O.OrderType='System' Then 'Las Vegas'
    When O.Location Like '%Houston%' Then 'Houston'
    WHen O.Location Like '%Orlando%' Then 'Orlando'
    When O.Location Like 'Penn%' Then 'Pennsylvania'
    When O.Location='Enterprises' Then 'Entertainment'
    Else O.Location
End