Tuning a Recursive Query

  • I have th e following recursive query:

    DECLARE@ConsultantIDASnVarChar(50)

    ,@PeriodDateASDateTime

    SET @ConsultantID = '0000344'

    SET @PeriodDate = '12-01-2007'

    -- Declare Local Variables

    Declare @MonthStartDt As DateTime

    DECLARE @MonthEndDt AS DateTime

    DECLARE @QuarterStartDtASDateTime

    DECLARE @QuarterEndDt AS DateTime

    Declare @YearASDateTime

    Declare @PeriodStartDt As DateTime

    Declare @PeriodEndDt as DateTime

    -- Breakdown @PeriodDate into Quarter and Months

    SET @QuarterStartDt = DATEADD(quarter, DATEDIFF(quarter, 0, @periodDate), 0)

    Set @QuarterEndDt = DATEADD(quarter, DATEDIFF(quarter, -1, @PeriodDate), -1)

    SET @Year = DATEADD(yy, DATEDIFF(yy,0,@PeriodStartDt), 0)

    SET @MonthStartDt = DATEADD(month, DATEDIFF(month, 0, @PeriodDate), 0)

    Set @MonthEndDt = DATEADD(month, DATEDIFF(month, -1, @PeriodDate), -1)

    IF @MonthEndDt > GETDATE()

    BEGIN

    SET @MonthEndDt = GETDATE()

    END;

    -- Declare the Downline Recursive Query

    With downline (ConsultantID,EffectiveDate, ConsultantName,SponsorID,SponsorName,DownLineLevel,

    ConsultantXId,SponsorXID,Active, DeactivationDate,BumpupDate,CurrentLevelAchieveDate, CurrentLevelXID )

    AS

    (

    -- Anchor member defintion

    SELECTA.ConsultantID

    ,A.EffectiveDate

    ,A.FirstName + ' ' + A.LastName as ConsultantName

    ,CAST(A.SponsorID AS Nvarchar(MAX))

    ,'' As SponsorName

    ,0 as DownLineLevel

    ,A.ConsultantXID

    ,A.SponsorXID

    ,A.Active

    ,A.DeactivationDate

    ,A.BumpupDate

    ,A.CurrentLevelAchieveDate

    ,A.CurrentLevelXID

    FROM dbo.uvwConsultantDownLine A with (nolock)

    WHERE A.ConsultantID = @ConsultantID

    AND @MonthEndDt Between a.EffectiveDate and A.EffectiveEndDate -- 1 Record

    UNION ALL

    --Recursive member definition

    SELECTA.ConsultantID

    ,A.EffectiveDate

    ,A.FirstName + ' ' + A.LastName as ConsultantName

    ,cast(A.SponsorID as nvarchar(MAX)) AS SponsorID

    ,'' AS SponsorName

    ,DownLineLevel + 1

    ,A.ConsultantXID

    ,A.SponsorXID

    ,A.Active

    ,A.DeactivationDate

    ,A.BumpupDate

    ,A.CurrentLevelAchieveDate

    ,A.CurrentLevelXID

    FROM dbo.uvwConsultantDownLine AS A with (nolock)

    INNER JOIN DownLine AS B ON

    A.SponsorID = B.ConsultantID

    WHERE @MonthEndDt Between a.EffectiveDate and A.EffectiveEndDate

    ) -- Appropriate records

    -- Create the Temp table #Downline that returns the CTE results

    SELECT A.DownLineLevel

    -- ,C.BumpUpDate

    ,A.ConsultantID

    ,A.EffectiveDate

    ,UPPER(RTRIM(A.ConsultantName)) AS ConsultantName

    --,C.EmailAddress

    ,D.Title AS AchievedTitle

    , CONVERT(NVARCHAR(MAX),A.SponsorID) AS SponsorID

    ,A.SponsorName

    ,A.ConsultantXID

    ,A.SponsorXID

    ,A.Active

    ,A.DeactivationDate

    ,A.BumpupDate

    ,A.CurrentLevelAchieveDate

    ,A.CurrentLevelXID

    INTO #Downline

    FROM DownLine AS A with (noLock)

    LEFT OUTER JOIN SharedDimension.dbo.DimConsultantTitle AS D ON

    A.CurrentLevelXID = D.XID

    WHERE DownLineLevel <= 4

    GROUP BY A.ConsultantID

    ,A.EffectiveDate

    ,A.ConsultantName

    ,A.SponsorID

    ,A.SponsorName

    ,DownLineLevel

    --,C.BumpUpDate

    --,C.EmailAddress

    ,D.Title

    ,A.ConsultantXID

    ,A.SponsorXID

    ,A.Active

    ,A.DeactivationDate

    ,A.BumpupDate

    ,A.CurrentLevelAchieveDate

    ,A.CurrentLevelXID

    Select D.ConsultantID ,D.Downlinelevel

    ,D.ConsultantName

    ,D.SponsorName

    ,D.EffectiveDate

    ,D.SponsorID

    ,D.AchievedTitle

    ,ISNULL(CONVERT(CHAR(10),D.CurrentLevelAchieveDate,101),'') AS AchieveDate

    ,ISNULL(Convert(Char(10),D.BumpupDate,101),'') as BumpupDate

    ,ISNULL(Sum(Case

    WHEN O.OrderCreateDate Between @QuarterStartDt And @QuarterEndDt THEN O.PartOneTotal

    ELSE 0

    END),0) AS QuarterToDate_total

    ,ISNULL(Sum(Case

    WHEN O.OrderCreateDate Between @MonthStartDt And @MonthEndDt THEN O.PartOneTotal

    ELSE 0

    END),0) AS MonthToDate_Total

    ,D.ConsultantXID

    ,D.SponsorXID

    ,ISNULL((SELECT ISNULL(SUM(v.PartOneTotal),0) FROM uvw_DownlineOrder V

    INNER JOIN SharedDimension.dbo.DimOrderType AS T ON

    V.OrderTypeXID = T.XID

    WHERE (T.OrderType NOT In ('Credit'))

    AND D.ConsultantID = V.ConsultantID

    AND V.OrderCreateDate Between @QuarterStartDt And @QuarterEndDt),0) AS QuarterToDate_Volume

    ,ISNULL((SELECT ISNULL(SUM(v.PartOneTotal),0) FROM uvw_DownlineOrder V

    INNER JOIN SharedDimension.dbo.DimOrderType AS T ON

    V.OrderTypeXID = T.XID

    WHERE (T.OrderType NOT In ('Credit'))

    AND D.ConsultantID = V.ConsultantID

    AND V.OrderCreateDate Between @MonthStartDt And @MonthEndDt),0) AS MonthToDate_Volume

    ,D.Active

    --,ISNULL((SELECT DISTINCT repFlag FROM dbo.udfGetRepromotes (@ConsultantID) r WHERE repFlag ='X'

    --AND d.ConsultantID = r.Consultantid ),' ') AS RepFlag

    ,ISNULL(r.RepFlag,' ')AS RepFlag

    from #Downline D with (nolock)

    LEFT OUTER JOIN uvw_DownlineOrder O ON D.ConsultantID = O.ConsultantID

    LEFT Outer JOIN Repromotes r ON d.ConsultantID = r.ConsultantID AND r.repflag = 'X'

    WHERE (D.Active = 1) OR (D.DeactivationDate BETWEEN @MonthStartDt AND @MonthEndDt)

    AND r.AchieveLevel >= 4

    GROUP BY D.ConsultantID ,D.Downlinelevel

    ,D.ConsultantName

    ,D.SponsorName

    ,D.EffectiveDate

    ,D.SponsorID

    ,D.AchievedTitle

    ,D.CurrentLevelAchieveDate

    ,D.BumpupDate

    ,D.ConsultantXID

    ,D.SponsorXID

    ,D.Active

    ,r.RepFlag

    -- Drop the temp table

    DROP TABLE #Downline

    It does what it's suppose to do but it really slow, any tips on how to tune it?

    Thanks,
    Art
    Database Analyst
    Tastefully Simple, Inc.

    alorenzini@tastefullysimple.com
    " Some days, it's not even worth chewing through the restraints!

  • Before you run the query, click 'Include actual execution plan' and see how it looks like when the query is finished. look for table scans for example, try to eliminate them (if number of rows in table is big) by indexing.

    that's for the beginning 😉

    Piotr

    ...and your only reply is slàinte mhath

  • You are so correct, I looked at the query plan and found I was missing few indexes, which I appl;ied and now its a speedy proc.

    Thanks

    Thanks,
    Art
    Database Analyst
    Tastefully Simple, Inc.

    alorenzini@tastefullysimple.com
    " Some days, it's not even worth chewing through the restraints!

Viewing 3 posts - 1 through 2 (of 2 total)

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