table scan problem - optimization of nonclustered indexes

  • Hi guys hope someone can help.

    I have a query that gets data to be inserted in a datawarehouse. The query is doing a table scan on the main table. I have tried quite a few combinations and diff indexes. The main table does not have a clustered index (problem 1). It does a sort that takes 83% and this is what I need to eliminate.

    Current indexes on T1 (Main Table) -

    Pk idx1 col's: Binary_key (Nonclustered)

    This is the structure of the query

    Select

    Case when T1.Address_Number = 0 AND T1.Document_Type in ('JI','JO','T4','TE') THEN Cast(T1.Company As Float) ELSE T1.Address_Number END As BBBEE_Address_Number,

    1 As Counter,

    T1.Document_Type,

    T1.GL_Offset,

    T1.Supplier_Invoice_Number,

    IsNull(T18.Date_BK,1) As Invoice_Date,

    T1.Account_ID ,

    Case when isnumeric(substring(serial_Number,2,24)) = 0 OR len(serial_Number) = 0 THEN 0 else Cast(substring(Serial_Number,2,len(Serial_Number)) As int) END As Asset_Item_Number,

    Isnull( T7.Time_BK ,1 ) as Time_BK,

    IsNull(T8.Business_Unit_BK,1) as Business_Unit_BK,

    T1.Ledger_Type as Ledger_Type,

    T1.Company as Company,

    T1.Address_Number as Address_Number,

    Case when isnumeric(T1.Sub_Ledger )= 0 OR T1.Sub_ledger_type != 'W' then 0 else Cast(T1.Sub_Ledger As Float) end as Sub_Ledger,

    IsNull(T16.Date_BK,1) as GL_Date,

    IsNull(T17.Date_BK ,1) as Batch_Date,

    T1.Reconciled_code,

    T1.Batch_Type,

    SUM(T1.Amount) As Amount,

    SUM(T1.Calculated_AC) As Calculated_AC,

    getDate() As modify_Date_Stamp

    From

    T1 with (nolock)

    Left Outer Join

    T2

    on

    T1.company = T2.Company

    Left Outer Join

    T7

    on

    T7.Fiscal_Pattern_BK = T2.Fiscal_Date_Pattern and

    T7.Fiscal_Period_End_Date = T1.Period_End_Date and

    T7.Fiscal_Period = T1.Period_Number

    Left Outer Join

    T8

    on

    LTrim(RTrim(T8.Business_Unit_Code)) = LTrim(RTrim(T1.Business_Unit))

    Left Outer Join

    T16

    on

    T1.Capture_Date = T16.Calendar_Date AND

    T2.Fiscal_Date_Pattern = T16.Fiscal_Pattern_BK

    Left Outer Join

    T17

    on

    T1.Batch_Date = T17.Calendar_Date AND

    T2.Fiscal_Date_Pattern = T17.Fiscal_Pattern_BK

    Left Outer Join

    T18

    on

    T1.Invoice_Date = T18.Calendar_Date AND

    T2.Fiscal_Date_Pattern = T18.Fiscal_Pattern_BK

    Where

    T1.GL_Posted_code = 'p' AND

    T7.Fiscal_Year > 2007 AND

    Cast(T1.Account_Id As int) > 2394039

    Group by

    Case when T1.Address_Number = 0 AND T1.Document_Type in ('JI','JO','T4','TE') THEN Cast(T1.Company As Float) ELSE T1.Address_Number END,

    T1.Document_Type,

    T1.GL_Offset,

    T1.Supplier_Invoice_Number,

    IsNull(T18.Date_BK,1),

    T1.Account_ID ,

    Case when isnumeric(substring(serial_Number,2,24)) = 0 OR len(serial_Number) = 0 THEN 0 else Cast(substring(Serial_Number,2,len(Serial_Number)) As int) END,

    Isnull( T7.Time_BK ,1 ) ,

    IsNull(T8.Business_Unit_BK,1) ,

    T1.Ledger_Type ,

    T1.Company ,

    T1.Address_Number ,

    Case when isnumeric(T1.Sub_Ledger )= 0 OR T1.Sub_ledger_type != 'W' then 0 else Cast(T1.Sub_Ledger As Float) end ,

    IsNull(T16.Date_BK,1) ,

    IsNull(T17.Date_BK ,1) ,

    T1.Reconciled_code,

    T1.Batch_Type

    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
    Do not reinvent the wheel.
    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

  • Please post table definitions, index definitions and execution plan, as per http://www.sqlservercentral.com/articles/SQLServerCentral/66909/

    Gail Shaw
    Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
    SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability

    We walk in the dark places no others will enter
    We stand on the bridge and no one may pass

Viewing 2 posts - 1 through 1 (of 1 total)

You must be logged in to reply to this topic. Login to reply