Writing SQL Server Triggers Part 1:  After Triggers

Data Manipulation Language (DML) triggers are special types of stored procedures that operate on tables and views in a SQL Server database.  These triggers are invoked when data is inserted, updated, or deleted.  There are two potential scopes for these DML triggers; INSTEAD-OF the operation or AFTER the operation.  INSTEAD-OF triggers are typically used with views and happen (you guessed it) instead of the actual INSERT, UPDATE, or DELETE operation.  AFTER triggers, what we'll be focusing on today, occur after the operation has occurred.

It is worth noting that of this writing, there is no such thing as a “BEFORE trigger” in the SQL Server world.  It is my hopes that this is remedied in a future release.

When an operation occurs on a table with an AFTER trigger attach to it, such as an INSERT statement, the following actions occur:

1.      Constraints are checked to ensure that the INSERT does not violate any defined rules.
2.      The record is added to the table.
3.      The trigger is invoked.

The AFTER trigger always occurs AFTER the original DML statement occurs.  However, these triggers still execute in the context of the original transaction.  So, you can roll these transactions back in your AFTER trigger but only after all of the work has been done by the DML statement.

DML triggers have access to two special tables that store a copy of the data involved in the DML operation.  Below is an outline of these tables.

INSERTED - this pseudo table holds values for INSERT and UPDATE operations.  The structure of this table ALWAYS mimics the table where the data was modified.  For INSERT operations, this table holds the newly inserted values.  For UPDATE operations, this table holds the new set of values: the values that the field(s) were updated to.  

DELETED - this pseudo table holds values for UPDATE and DELETE operations.  The structure of this table ALWAYS mimics the table where the data was modified.  For UPDATE operations, this table holds the old set of values: the values that the columns in the table held BEFORE the update occurred.  For DELETE operations, this table holds the set of record(s) that were deleted.

Another MAJOR point with triggers is that the data you have access to in the trigger occurs at the BATCH level, and never EVER at the row level.  There is absolutely no built-in functionality in SQL Server to handle records in a trigger row-by-row.  To see how you can handle these situations, see my article here.  

So, lets take a look at how the INSERTED and DELETED tables are used when a DML statement occurs inside an AFTER trigger.  

First, I'll create a table and insert some data into it.
 

CREATE TABLE TestTable
                      (
                      	ID INT IDENTITY(1,1),
                      	ValueField INT DEFAULT(1),
                      	NewGuid UNIQUEIDENTIFIER DEFAULT(NEWID())
                      )
                      GO
                      INSERT INTO TestTable DEFAULT VALUES
                      GO 10
                      GO
                      

Select all Open in new window

Now I'll define an AFTER trigger on the TestTable table.
This trigger will be invoked whenever an INSERT, UPDATE, or DELETE occurs and simply outputs the values from the INSERTED and DELETED tables to the screen.
 

CREATE TRIGGER TR_TestTable_IUD
                      ON TestTable
                      AFTER INSERT, UPDATE, DELETE
                      AS
                      BEGIN
                      	SELECT 'INSERTED' AS TableName, * FROM INSERTED
                      	SELECT 'DELETED' AS TableName, * FROM DELETED
                      END
                      GO
                      

Select all Open in new window

Now that I have my trigger defined, I want to test it.
In the script below I'll insert a single value.  
 

INSERT INTO TestTable DEFAULT VALUES
                      

Select all Open in new window

My output from this statement is below.  
      Notice that only the INSERTED table contains values.  This is due to the fact that for an INSERT operation, this table contains the newly inserted values into the table.

Now I'll delete a single record from the table.  
(Note:  the following syntax for the DELETE and UPDATE operations will only work for SQL Server 2005 and later.)
 

DELETE TOP(1) FROM TestTable
                      GO
                      

Select all Open in new window

The output from the DELETE statement is below.  
  Notice this time that only the DELETED table contains data.

Finally, I'll run an UPDATE statement on the table to update a single record.  
 

UPDATE TOP(1) t
                      SET ValueField = ValueField + 1
                      FROM TestTable t
                      

Select all Open in new window

And, here's the output.  
  Notice this time that the INSERTED and DELETED tables both contain data.  The INSERTED table contains the NEW values; the DELETED table, the OLD values.

So, that's how you can define a basic SQL Server AFTER trigger.  To recap: we took a look at how you can define the trigger structure, how you can define the DML events that will invoke the trigger, and how the INSERTED and DELETE pseudo tables are populated based on the operation type.

Happy Coding!
Tim