SQL Server - synchronizing 2 tables on 2 different

问题:

I have 2 tables with same schema on 2 different databases on the same server with SQL Server 2008 R2. One table gets updated with data more often.

Now there is a need to keep these 2 table in sync. This can happen as a nightly process. What is the best methodology to achieve the sync. process ?

回答1:

Using MERGE is your best bet. You can control each of the conditions. WHEN MATCHED THEN, WHEN UNMATCHED THEN etc.

MERGE - Technet

MERGE- MSDN (GOOD!)

Example A: Transactional usage - Table Variables - NO

DECLARE @Source TABLE (ID INT)
DECLARE @Target TABLE (ID INT)

INSERT INTO @Source (ID) VALUES (1),(2),(3),(4),(5)

BEGIN TRANSACTION

MERGE @Target AS T
USING @Source AS S
ON (S.ID = T.ID)
WHEN NOT MATCHED THEN
    INSERT (ID) VALUES (S.ID);

ROLLBACK TRANSACTION
SELECT  'FAIL' AS Test,*
FROM    @Target

Example B: Transactional usage - Physical Tables

CREATE TABLE SRC (ID INT);
CREATE TABLE TRG (ID INT);

INSERT INTO SRC (ID) VALUES (1),(2),(3),(4),(5)

BEGIN TRANSACTION

MERGE TRG AS T
USING SRC AS S
ON (S.ID = T.ID)
WHEN NOT MATCHED THEN
    INSERT (ID) VALUES (S.ID);

ROLLBACK TRANSACTION
SELECT  'FAIL' AS Test,*
FROM    TRG

Example C: Transactional usage - Tempdb (local & global)

CREATE TABLE #SRC (ID INT);
CREATE TABLE #TRG (ID INT);

INSERT INTO #SRC (ID) VALUES (1),(2),(3),(4),(5)

BEGIN TRANSACTION

MERGE #TRG AS T
USING #SRC AS S
ON (S.ID = T.ID)
WHEN NOT MATCHED THEN
    INSERT (ID) VALUES (S.ID);

ROLLBACK TRANSACTION
SELECT  'FAIL' AS Test,*
FROM    #TRG

回答2:

You probably can use sql server's tablediff.exe command line utility. It can do table-by-table, one-off compare between two tables and generate the sql automatically for you to sync the dest to the source.

There's also a GUI wrapper around it http://code.google.com/p/sqltablediff/ which makes the job even easier. It will generate the command line for you.

You can then create a scheduled task to run the command line, and then execute the generated sql scripts.

回答3:

You can select from the different databases and use a cursor to loop the selected data. Within that cursor you can do some logic and update or delete from the target table.

Also SQL 2008 has a nice new MERGE statement which you can use to select/insert/update in one T-SQL query. http://technet.microsoft.com/en-us/library/bb510625%28v=sql.105%29.aspx

For more complex processes i use the first option. For more straight forward sync tasks i use the second option.

As an extra option there is also Server Integration Services (SSIS): http://blogs.msdn.com/b/jorgepc/archive/2010/12/07/synchronize-two-tables-using-sql-server-integration-services-ssis-part-i-of-ii.aspx