Generate SQL MERGE statements with Table data
This system stored procedure takes a table name as a parameter and generates a
MERGE statement containing all the table data.
This is useful if you need to migrate static data between databases, eg. the generated MERGE statement can be included in source control and used to deploy data between DEV/TEST/PROD.
The stored procedure itself is installed within the
[master] database as a system object, allowing the proc to be called within the context of user databases (e.g.
EXEC Northwind.dbo.sp_generate_merge 'Region')
- Include or exclude specific columns from output (eg. exclude DateCreated/DateModified columns)
- Only update the target database when changes in the source data are found
- Support for larger tables (gets around character limitations in some SQL clients)
How Does it Work?
The generated MERGE statement populates the target table to match the source data. This includes the removal of any excess rows that are not present in the source.
When the generated MERGE statement is executed, the following logic is applied based on whether a match is found:
- If the source row does not exist in the target table, an
- If a given row in the target table does not exist in the source, a
- If the source row already exists in the target table and has changed, an
- If the source row already exists in the target table but the data has not changed, no action is performed (configurable)
- Generate statements for static data tables, store the .SQL file in source control/add it to a Visual Studio Database Project and use it as part of your Dev/Test/Prod deployments. The generated statements are re-runnable, so you can make changes to the file and easily migrate those changes between environments.
- Generate statements from your Production tables and then run those statements in your Dev/Test environments. Schedule this as part of a SQL Job to keep all of your environments in-sync.
- Enter test data into your Dev environment, and then generate statements from the Dev tables so that you can always reproduce your test database with valid sample data.
This procedure was adapted from sp_generate_inserts, written by Narayana Vyas Kondreddi (http://vyaskn.tripod.com). I made a number of attempts to get in touch with Vyas but unfortunately have not been able to reach him. No copyright infringement is intended and I will of course respect his wishes if asks for this to be removed.
I would also like to acknowledge:
- Bill Graziano -- Blog post that provided the groundwork for MERGE statement generation (http://weblogs.sqlteam.com/billg/archive/2011/02/15/generate-merge-statements-from-a-table.aspx)
- Bill Gibson -- Blog post that detailed the static data table use case; the inspiration for this proc (http://blogs.msdn.com/b/ssdt/archive/2012/02/02/including-data-in-an-sql-server-database-project.aspx)
- Nathan Skerl -- Provided a novel way of working around the 8000 character limit in SSMS (http://stackoverflow.com/a/10489767/266882)
Simply execute the script, which will install it in
[master] database as a system procedure (making it executable within user databases).
This procedure has explicit support for the following datatypes: (small)datetime(2), datetimeoffset, (n)varchar, (n)text, (n)char, xml, int, float, real, (small)money, timestamp, rowversion, uniqueidentifier, (var)binary, hierarchyid, geometry and geography. All others are implicitly converted to their CHAR representations so YMMV depending on the datatype.
The Image datatype is not supported and an error will be thrown if these are not excluded using the
When using the
@hash_compare_column parameter, all columns in the source and target table must be implicitly convertible to strings (due to the use of
CONCAT in the proc to calculate the hash value). This means that the following data types are not supported with
@hash_compare_column: xml, hierarchyid, geometry and geography.
- Ensure that your SQL client is configured to send results to grid.
- Execute the proc, providing the source table name as a parameter
- Click the hyperlink within the resultset.
- Copy the SQL (excluding the Output tags) and paste into a new query window to execute.
To generate a MERGE statement containing all data within the Person.AddressType table, excluding the ModifiedDate and rowguid columns:
EXEC AdventureWorks.dbo.sp_generate_merge @schema = 'Person', @table_name ='AddressType', @cols_to_exclude = '''ModifiedDate'',''rowguid'''
SET NOCOUNT ON GO SET IDENTITY_INSERT [Person].[AddressType] ON GO MERGE INTO [Person].[AddressType] AS Target USING (VALUES (1,'Billing') ,(2,'Home') ,(3,'Main Office') ,(4,'Primary') ,(5,'Shipping') ,(6,'Contact') ) AS Source ([AddressTypeID],[Name]) ON (Target.[AddressTypeID] = Source.[AddressTypeID]) WHEN MATCHED AND ( NULLIF(Source.[Name], Target.[Name]) IS NOT NULL OR NULLIF(Target.[Name], Source.[Name]) IS NOT NULL) THEN UPDATE SET [Name] = Source.[Name] WHEN NOT MATCHED BY TARGET THEN INSERT([AddressTypeID],[Name]) VALUES(Source.[AddressTypeID],Source.[Name]) WHEN NOT MATCHED BY SOURCE THEN DELETE; SET IDENTITY_INSERT [Person].[AddressType] OFF GO SET NOCOUNT OFF GO