SQL server 触发器，在触发Merge过程中，逐行触发的解决办法 用group by 避免是一次触发中的多行更新或删除。

When I share "tsql trigger row by row" on bing.

I got :点击打开链接

Trigger:

IF OBJECT_ID('dbo.UPD_ORDERTRACKING_WEBE','TR') IS NOT NULL
BEGIN
DROP TRIGGER dbo.UPD_ORDERTRACKING_WEBE
END
GO
CREATE TRIGGER  dbo.UPD_ORDERTRACKING_WEBE
ON dbo.avORDERTRACKING
AFTER  INSERT,UPDATE
AS
BEGIN
/*SET NOCOUNT ON added to prevent extra result sets from
*interfering with SELECT statements.*/
SET NOCOUNT ON;
--organize source data
;WITH s AS (
SELECT INSERTED.dataareaid AS dataareaid,INSERTED.CustAccount AS accountnum,
GETDATE() AS refresh_ts,MAX(INSERTED.recid) recid,MAX(INSERTED.RECVERSION)
RECVERSION,MAX(INSERTED.PARTITION) PARTITION FROM
INSERTED WHERE  INSERTED.avorderstatus <> 'DELETED'
GROUP BY INSERTED.dataareaid,INSERTED.CustAccount
--for records of this set maybe not unique,using 'group by' can prevent updating or deleting the same row twice.
)
--begin merge AVCUSTTABLE_WEBE_LOG
MERGE INTO AVCUSTTABLE_WEBE_LOG d USING s
ON d.dataareaid = s.dataareaid AND d.accountnum = s.accountnum
WHEN matched THEN
UPDATE SET    d.refresh_ts = s.refresh_ts
WHEN NOT matched THEN
INSERT ( DATAAREAID,ACCOUNTNUM,REFRESH_TS,RECID,PARTITION,RECVERSION)VALUES
( s.DATAAREAID,s.ACCOUNTNUM,s.REFRESH_TS,s.recid,S.PARTITION,s.RECVERSION)
;--end merge
END
GO

My test case:

1. Inserting 4111 records,it costs zero second.

Insert records into avordertrancking.



Check the statistics, it costs 2+35+37=74ms to insert 4111 records into avordertrancking.



Check the result set, it has inserted.



2. Updating 4111 records,it costs zero ms

Update all records in avordertrancking. it costs 7+18+11=36ms



Check the result set ,it has updated



3. Additional It raise error without grouping

IF OBJECT_ID('dbo.UPD_ORDERTRACKING_WEBE','TR') IS NOT NULL
BEGIN
DROP TRIGGER dbo.UPD_ORDERTRACKING_WEBE
END
GO
CREATE TRIGGER  dbo.UPD_ORDERTRACKING_WEBE
ON dbo.avORDERTRACKING
AFTER  INSERT,UPDATE
AS
BEGIN
/*SET NOCOUNT ON added to prevent extra result sets from
*interfering with SELECT statements.*/
SET NOCOUNT ON;
--organize source data
;WITH s AS (
SELECT INSERTED.dataareaid AS dataareaid,INSERTED.CustAccount AS accountnum,
GETDATE() AS refresh_ts, INSERTED.recid , INSERTED.RECVERSION
RECVERSION, INSERTED.PARTITION  FROM
INSERTED WHERE  INSERTED.avorderstatus <> 'DELETED'
--for records of this set maybe not unique,using 'group by' can prevent updating or deleting the same row twice.
)
--begin merge AVCUSTTABLE_WEBE_LOG
MERGE INTO AVCUSTTABLE_WEBE_LOG d USING s
ON d.dataareaid = s.dataareaid AND d.accountnum = s.accountnum
WHEN matched THEN
UPDATE SET    d.refresh_ts = s.refresh_ts
WHEN NOT matched THEN
INSERT ( DATAAREAID,ACCOUNTNUM,REFRESH_TS,RECID,PARTITION,RECVERSION)VALUES
( s.DATAAREAID,s.ACCOUNTNUM,s.REFRESH_TS,s.recid,S.PARTITION,s.RECVERSION)
;--end merge
END
GO

Here is example contains:

01create apoms database.sql
02dbo.AVCUSTTABLE_WEBE_LOG.Table.sql
03dbo.avORDERTRACKING_bak.Table.sql
04dbo.avordertracking.Table.sql
05BMSSA.UPD_ORDERTRACKING_WEBE.sql
06[APOMS].[dbo].[avORDERTRACKING_bak] 2014-10-29 11-23-36.sql
07test.sql
runall.bat

download:

点击打开链接