Microsoft SQL-Abfrage Server verbessern

Question

Gibt es eine Chance Geschwindigkeit der Abfrage unten zu verbessern:

select distinct 
    a.InvoiceAcc, 
    (select top 1 b.CustomerName 
    from dbo.tblsales as b 
    where b.InvoiceAcc = a.InvoiceAcc), 
    (select sum(b.SalesValue) 
    from dbo.tblsales as b 
    where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-1') as [01], 
    (select sum(b.SalesValue) 
    from dbo.tblsales as b 
    where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-2') as [02], 
    (select sum(b.SalesValue) 
    from dbo.tblsales as b 
    where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-3') as [03], 
    (select sum(b.SalesValue) 
    from dbo.tblsales as b 
    where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-4') as [04], 
    (select sum(b.SalesValue) 
    from dbo.tblsales as b 
    where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-5') as [05], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-6') as [06], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-7') as [07], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-8') as [08], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-9') as [09], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-10') as [10], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-11') as [11], 
    (select sum(b.SalesValue) from dbo.tblsales as b where b.InvoiceAcc = a.InvoiceAcc and b.Month = '2016-12') as [12] 
from 
    dbo.tblsales as a 
group by 
    a.InvoiceAcc, CustomerName 

Was würde die Ansicht in SQL Server mit params Monat gesendet, wie 2016 und InvoiceAcc (nicht erforderlich)?

Vielen Dank.

Answer 1

Basierend auf Ihrem Beispiel wollen Sie diese

select 
    InvoiceAcc, 
    CustomerName, 
    sum(CASE WHEN [Month] = '2016-1' then salesvalue else 0 end) as [01], 
    sum(CASE WHEN [Month] = '2016-2' then salesvalue else 0 end) as [02], 
    sum(CASE WHEN [Month] = '2016-3' then salesvalue else 0 end) as [03], 
    sum(CASE WHEN [Month] = '2016-4' then salesvalue else 0 end) as [04], 
    sum(CASE WHEN [Month] = '2016-5' then salesvalue else 0 end) as [05], 
    sum(CASE WHEN [Month] = '2016-6' then salesvalue else 0 end) as [06], 
    sum(CASE WHEN [Month] = '2016-7' then salesvalue else 0 end) as [07], 
    sum(CASE WHEN [Month] = '2016-8' then salesvalue else 0 end) as [08], 
    sum(CASE WHEN [Month] = '2016-9' then salesvalue else 0 end) as [09], 
    sum(CASE WHEN [Month] = '2016-10' then salesvalue else 0 end) as [10], 
    sum(CASE WHEN [Month] = '2016-11' then salesvalue else 0 end) as [11], 
    sum(CASE WHEN [Month] = '2016-12' then salesvalue else 0 end) as [12] 
from dbo.tblsales 
group by InvoiceAcc, CustomerName 

ich den Test nicht finden (zB Month = '2016-1') fremd zu sein - ich ein Datumsfeld erwarten würde und ein Test wie dieser MONTH(invoiceDate) = 1

ich auch denke, dass die Verwendung von distinct in sql oft ein Hinweis auf eine schlechte Abfrage ist - es ist selten, dass eine korrekt entworfene Abfrage von einem gut entworfenen Modell eindeutig ist.

---

Wenn Sie die beste Art und Weise zusätzliche Namen entfernen möchten, es zu tun, ist dies:

select 
    InvoiceAcc, 
    CustomerName, 
    sum(CASE WHEN [Month] = '2016-1' then salesvalue else 0 end) as [01], 
    sum(CASE WHEN [Month] = '2016-2' then salesvalue else 0 end) as [02], 
    sum(CASE WHEN [Month] = '2016-3' then salesvalue else 0 end) as [03], 
    sum(CASE WHEN [Month] = '2016-4' then salesvalue else 0 end) as [04], 
    sum(CASE WHEN [Month] = '2016-5' then salesvalue else 0 end) as [05], 
    sum(CASE WHEN [Month] = '2016-6' then salesvalue else 0 end) as [06], 
    sum(CASE WHEN [Month] = '2016-7' then salesvalue else 0 end) as [07], 
    sum(CASE WHEN [Month] = '2016-8' then salesvalue else 0 end) as [08], 
    sum(CASE WHEN [Month] = '2016-9' then salesvalue else 0 end) as [09], 
    sum(CASE WHEN [Month] = '2016-10' then salesvalue else 0 end) as [10], 
    sum(CASE WHEN [Month] = '2016-11' then salesvalue else 0 end) as [11], 
    sum(CASE WHEN [Month] = '2016-12' then salesvalue else 0 end) as [12] 
from (
    SELECT InvoiceAcc, CustomerName, Month, salesvalue, 
     ROW_NUMBER() OVER (PARTITION BY InvoiceAcc ORDER BY CustomerName) AS RN 
     FROM tblsales 
) x 
WHERE RN=1 
group by InvoiceAcc, CustomerName 

Answer 2

Sie könnten versuchen, Pivot für die Leistung Summe besser

select 
InvoiceAcc, 
CustomerName, 
[2016-1] as [01], 
[2016-2] as [02], 
[2016-3] as [03], 
[2016-4] as [04], 
[2016-5] as [05], 
[2016-6] as [06], 
[2016-7] as [07], 
[2016-8] as [08], 
[2016-9] as [09], 
[2016-10] as [10], 
[2016-11] as [11], 
[2016-12] as [12] 
from (
select 
InvoiceAcc, 
CustomerName, 
[Month], 
sum(SalesValue) SalesValue 
from dbo.tblsales 
group by InvoiceAcc, CustomerName,[Month]) tb 
pivot 
(
sum(SalesValue) 
for [Month] in ([2016-1],[2016-2] ,[2016-3],[2016-4],[2016-5],[2016-6],[2016-7],[2016-8],[2016-9],[2016-10],[2016-11],[2016-12]) 
) pvt