Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Welcome To Ask or Share your Answers For Others

Categories

0 votes
283 views
in Technique[技术] by (71.8m points)

sql server - Handling multiple records in a MS SQL trigger

I am having to use triggers in MSSQL for the first time, well triggers in general. Having read around and tested this myself I realise now that a trigger fires per command and not per row inserted, deleted or updated.

The entire thing is some statistics for an advertising system. Our main stat table is rather large and doesn't contain the data in a way that makes sense in most cases. It contains one row per advert clicked, viewed and etc. As a user one is more inclined to want to view this as day X has Y amount of clicks and Z amount of views and so forth. We have done this purely based on a SQL query so far, getting this sort of report from the main table, but as the table has grown so does the time for that query to execute. Because of this we have opted for using triggers to keep another table updated and hence making this a bit easier on the SQL server.

My issue is now to get this working with multiple records. What I have done is to create 2 stored procedures, one for handling the operation of an insert, and one for a delete. My insert trigger (written to work with a single record) then graps the data off the Inserted table, and sends it off to the stored procedure. The delete trigger works in the same way, and (obviously?) the update trigger does the same as a delete + an insert.

My issue is now how to best do this with multiple records. I have tried using a cursor, but as far as I have been able to read and see myself, this performs really badly. I have considered writing some "checks" as well - as in checking to see IF there are multiple records in the commands and then go with the cursor, and otherwise simply just avoid this. Anyhow, here's my solution with a cursor, and im wondering if there's a way of doing this better?

CREATE TRIGGER [dbo].[TR_STAT_INSERT]
   ON  [iqdev].[dbo].[Stat]
   AFTER INSERT
AS 
BEGIN
    SET NOCOUNT ON;

    DECLARE @Date DATE 
    DECLARE @CampaignId BIGINT
    DECLARE @CampaignName varchar(500)
    DECLARE @AdvertiserId BIGINT
    DECLARE @PublisherId BIGINT
    DECLARE @Unique BIT
    DECLARE @Approved BIT
    DECLARE @PublisherEarning money
    DECLARE @AdvertiserCost money
    DECLARE @Type smallint

    DECLARE InsertCursor CURSOR FOR SELECT Id FROM Inserted
    DECLARE @curId bigint

    OPEN InsertCursor

    FETCH NEXT FROM InsertCursor INTO @curId

    WHILE @@FETCH_STATUS = 0
    BEGIN

        SELECT @Date = [Date], @PublisherId = [PublisherCustomerId], @Approved = [Approved], @Unique = [Unique], @Type = [Type], @AdvertiserCost = AdvertiserCost, @PublisherEarning = PublisherEarning
        FROM Inserted
        WHERE Id = @curId

        SELECT @CampaignId = T1.CampaignId, @CampaignName = T2.Name, @AdvertiserId = T2.CustomerId
        FROM Advert AS T1
        INNER JOIN Campaign AS T2 on T1.CampaignId = T2.Id
        WHERE T1.Id = (SELECT AdvertId FROM Inserted WHERE Id = @curId)

        EXEC ProcStatInsertTrigger @Date, @CampaignId, @CampaignName, @AdvertiserId, @PublisherId, @Unique, @Approved, @PublisherEarning, @AdvertiserCost, @Type

        FETCH NEXT FROM InsertCursor INTO @curId
    END

    CLOSE InsertCursor
    DEALLOCATE InsertCursor
END

The stored procedure is rather big and intense and I do not think there's a way of having to avoid looping through the records of the Inserted table in one way or another (ok, maybe there is, but I'd like to be able to read the code too :p), so I'm not gonna bore you with that one (unless you like to think otherwise). So pretty much, is there a better way of doing this, and if so, how?

EDIT: Well after request, here's the sproc

CREATE PROCEDURE ProcStatInsertTrigger 
    @Date DATE,
    @CampaignId BIGINT,
    @CampaignName varchar(500),
    @AdvertiserId BIGINT,
    @PublisherId BIGINT,
    @Unique BIT,
    @Approved BIT,
    @PublisherEarning money,
    @AdvertiserCost money,
    @Type smallint
AS
BEGIN
    -- SET NOCOUNT ON added to prevent extra result sets from
    -- interfering with SELECT statements.
    SET NOCOUNT ON;
IF @Approved = 1
        BEGIN
            DECLARE @test bit

            SELECT @test = 1 FROM CachedStats WHERE [Date] = @Date AND CampaignId = @CampaignId AND CustomerId = @PublisherId

            IF @test IS NULL
                BEGIN
                    INSERT INTO CachedStats ([Date], CustomerId, CampaignId, CampaignName) VALUES (@Date, @PublisherId, @CampaignId, @CampaignName)
                END

            SELECT @test = NULL

                    DECLARE @Clicks int
                    DECLARE @TotalAdvertiserCost money
                    DECLARE @TotalPublisherEarning money
                    DECLARE @PublisherCPC money
                    DECLARE @AdvertiserCPC money

                    SELECT @Clicks = Clicks, @TotalAdvertiserCost = AdvertiserCost + @AdvertiserCost, @TotalPublisherEarning = PublisherEarning + @PublisherEarning FROM CachedStats
                    WHERE [Date] = @Date AND CustomerId = @PublisherId AND CampaignId = @CampaignId

                    IF @Type = 0 -- If click add one to the calculation
                        BEGIN
                            SELECT @Clicks = @Clicks + 1
                        END

                    IF @Clicks > 0
                        BEGIN
                            SELECT @PublisherCPC = @TotalPublisherEarning / @Clicks, @AdvertiserCPC = @TotalAdvertiserCost / @Clicks
                        END
                    ELSE
                        BEGIN
                            SELECT @PublisherCPC = 0, @AdvertiserCPC = 0
                        END
            IF @Type = 0
                BEGIN

                    UPDATE CachedStats SET
                        Clicks = @Clicks,
                        UniqueClicks = UniqueClicks + @Unique,
                        PublisherEarning = @TotalPublisherEarning,
                        AdvertiserCost = @TotalAdvertiserCost,
                        PublisherCPC = @PublisherCPC,
                        AdvertiserCPC = @AdvertiserCPC
                    WHERE [Date] = @Date AND CustomerId = @PublisherId AND CampaignId = @CampaignId
                END
            ELSE IF @Type = 1 OR  @Type = 4 -- lead or coreg
                BEGIN
                    UPDATE CachedStats SET
                        Leads = Leads + 1,
                        PublisherEarning = @TotalPublisherEarning,
                        AdvertiserCost = @TotalAdvertiserCost,
                        AdvertiserCPC = @AdvertiserCPC,
                        PublisherCPC = @AdvertiserCPC
                    WHERE [Date] = @Date AND CustomerId = @PublisherId AND CampaignId = @CampaignId
                END
            ELSE IF @Type = 3 -- Isale
                BEGIN
                    UPDATE CachedStats SET
                        Leads = Leads + 1,
                        PublisherEarning = @TotalPublisherEarning,
                        AdvertiserCost = @TotalAdvertiserCost,
                        AdvertiserCPC = @AdvertiserCPC,
                        PublisherCPC = @AdvertiserCPC,
                        AdvertiserOrderValue = @AdvertiserCost,
                        PublisherOrderValue = @PublisherEarning 
                    WHERE [Date] = @Date AND CustomerId = @PublisherId AND CampaignId = @CampaignId                 
                END
           ELSE IF @Type = 2 -- View
                BEGIN
                    UPDATE CachedStats SET
                        [Views] = [Views] + 1,
                        UniqueViews = UniqueViews + @Unique,
                        PublisherEarning = @TotalPublisherEarning,
                        AdvertiserCost = @TotalAdvertiserCost,
                        PublisherCPC = @PublisherCPC,
                        AdvertiserCPC = @AdvertiserCPC
                    WHERE [Date] = @Date AND CustomerId = @PublisherId AND CampaignId = @CampaignId         
                END
        END
END

After help, here's my final result, posted in case others have a similiar issue

CREATE TRIGGER [dbo].[TR_STAT_INSERT]
   ON  [iqdev].[dbo].[Stat]
   AFTER INSERT
AS 
BEGIN

    SET NOCOUNT ON

    -- insert all missing "CachedStats" rows
    INSERT INTO
        CachedStats ([Date], AdvertId, CustomerId, CampaignId, CampaignName) 
    SELECT DISTINCT
        CONVERT(Date, i.[Date]), i.AdvertId, i.[PublisherCustomerId], c.Id, c.Name
    FROM
        Inserted i
        INNER JOIN Advert AS   a ON a.Id = i.AdvertId
        INNER JOIN Campaign AS c ON c.Id = a.CampaignId
    WHERE
        i.[Approved] = 1
        AND NOT EXISTS (
                SELECT 1 
                FROM CachedStats as t
                WHERE 
                        [Date] = CONVERT(Date, i.[Date])
                        AND CampaignId = c.Id 
                        AND CustomerId = i.[PublisherCustomerId]
                        AND t.AdvertId = i.AdvertId
        )

  -- update all affected records at once
    UPDATE 
        CachedStats
    SET
        Clicks = 
            Clicks + (
                SELECT COUNT(*) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.AdvertId
                AND   s.[Type] = 0
            ),
        UniqueClicks = 
            UniqueClicks + (
                SELECT COUNT(*) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.[Unique] = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.AdvertId
                AND   s.[Type] = 0
            ),
        [Views] = 
            [Views] + (
                SELECT COUNT(*) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.AdvertId
                AND   s.[Type] = 2
            ),
        UniqueViews = 
            UniqueViews + (
                SELECT COUNT(*) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.[Unique] = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.AdvertId
                AND   s.[Type] = 2
            ),
        Leads = 
            Leads + (
                SELECT COUNT(*) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.[Unique] = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.AdvertId
                AND   s.[Type] IN (1,3,4)
            ),
        PublisherEarning =
            CachedStats.PublisherEarning + ISNULL((
                SELECT SUM(PublisherEarning) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.AdvertId           

            ), 0),
        AdvertiserCost =
            CachedStats.AdvertiserCost + ISNULL((
                SELECT SUM(AdvertiserCost) FROM Inserted s
                WHERE s.Approved = 1
                AND   s.PublisherCustomerId = i.PublisherCustomerId
                AND   CONVERT(Date, s.[Date]) = CONVERT(Date, i.[Date])
                AND   s.AdvertId = i.Ad

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome To Ask or Share your Answers For Others

1 Reply

0 votes
by (71.8m points)

The trick with these kinds of situations is to turn the sequential operation (for each record do xyz) into a set-based operation (an UPDATE statement).

I have analyzed your stored procedure and merged your separate UPDATE statements into a single one. This single statement can then be transformed into a version that can be applied to all inserted records at once, eliminating the need for a stored procedure and thereby the need for a cursor.

EDIT: Below is the code that we finally got working. Execution time for the whole operation went down from "virtually forever" (for the original solution) to something under one second, according to the OP's feedback. Overall code size also decreased quite noticeably.

CREATE TRIGGER [dbo].[TR_STAT_INSERT]
   ON  [iqdev].[dbo].[Stat]
   AFTER INSERT
AS 
BEGIN
  SET NOCOUNT ON

  -- insert all missing "CachedStats" rows
  INSERT INTO
    CachedStats ([Date], AdvertId, CustomerId, CampaignId, CampaignName) 
  SELECT DISTINCT
    CONVERT(Date, i.[Date]), i.AdvertId, i.PublisherCustomerId, c.Id, c.Name
  FROM
    Inserted i
    INNER JOIN Advert   a ON a.Id = i.AdvertId
    INNER JOIN Campaign c ON c.Id = a.CampaignId
  WHERE
    i.Approved = 1
    AND NOT EXISTS ( 
      SELECT 1 
      FROM   CachedStats
      WHERE  Advertid   = i.AdvertId AND
             CustomerId = i.PublisherCustomerId AND
             [Date]     = CONVERT(DATE, i.[Date])
    )

  -- update all affected records at once
  UPDATE 
    CachedStats
  SET
    Clicks               = Clicks               + i.AddedClicks,
    UniqueClicks         = UniqueClicks         + i.AddedUniqueClicks,
    [Views]              = [Views]              + i.AddedViews,
    UniqueViews          = UniqueViews          + i.AddedUniqueViews,
    Leads                = Leads                + i.AddedLeads,
    PublisherEarning     = PublisherEarning     + ISNULL(i.AddedPublisherEarning, 0),
    AdvertiserCost       = AdvertiserCost       + ISNULL(i.AddedAdvertiserCost, 0),
    PublisherOrderValue  = PublisherOrderValue  + ISNULL(i.AddedPublisherOrderValue, 0),
    AdvertiserOrderValue = AdvertiserOrderValue + ISNULL(i.AddedAdvertiserOrderValue, 0)
  FROM
    (
    SELECT
      AdvertId,
      CONVERT(DATE, [Date]) [Date],
      PublisherCustomerId,
      COUNT(*) NumRows,
      SUM(CASE WHEN Type IN (0)                      THEN 1 ELSE 0 END) AddedClicks,
      SUM(CASE WHEN Type IN (0)     AND [Unique] = 1 THEN 1 ELSE 0 END) AddedUniqueClicks,
      SUM(CASE WHEN Type IN (2)                      THEN 1 ELSE 0 END) AddedViews,
      SUM(CASE WHEN Type IN (2)     AND [Unique] = 1 THEN 1 ELSE 0 END) AddedUniqueViews,
      SUM(CASE WHEN Type IN (1,3,4) AND [Unique] = 1 THEN 1 ELSE 0 END) AddedLeads,
      SUM(PublisherEarning)                                      AddedPublisherEarning,
      SUM(AdvertiserCost)                                        AddedAdvertiserCost,
      SUM(CASE WHEN Type IN (3) THEN PublisherOrderValue  ELSE 0 END) AddedPublisherOrderValue,
      SUM(CASE WHEN Type IN (3) THEN AdvertiserOrderValue ELSE 0 END) AddedAdvertiserOrderValue
    FROM
      Inserted
    WHERE
      Approved = 1
    GROUP BY
      AdvertId,
      CONVERT(DATE, [Date]),
      PublisherCustomerId
    ) i 
    INNER JOIN CachedStats cs ON 
      cs.Advertid   = i.AdvertId AND
      cs.CustomerId = i.PublisherCustomerId AND
      cs.[Date]     = i.[Date]

  SET NOCOUNT OFF
END

The operations involving the CachedStats table will greatly benefit from one multiple-column index over (Advertid, CustomerId, [Date]) (as confirmed by the OP).


与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
OGeek|极客中国-欢迎来到极客的世界,一个免费开放的程序员编程交流平台!开放,进步,分享!让技术改变生活,让极客改变未来! Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Click Here to Ask a Question

...