Question

I have a large table in a database and I want to track changes made to the individual records. More precisely, I want to log the date and the changes made to the columns.

Since the table has 25+ comlumns I don’t want to test them individually.

The logging table looks like ID-Date-Table-Column-OldValue-NewValue

In my AFTER UPDATE trigger I’d like to check which columns have different values and log them into my table.

I know I can get the columns of the table with:

DECLARE @meta_table TABLE (  
    idx smallint Primary Key IDENTITY(1,1) 
    , TABLE_NAME nchar(100), COLUMN_NAME nchar(100), COLUMN_ID int 
) 

INSERT @meta_table 
SELECT TABLE_NAME, COLUMN_NAME,
    COLUMNPROPERTY(OBJECT_ID(TABLE_SCHEMA + '.' + TABLE_NAME),
    COLUMN_NAME, 'ColumnID') AS COLUMN_ID
FROM MYDATABASE.INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = 'myTable';

I can iterate the columns with:

SET @i = 1 
SET @numrows = (SELECT COUNT(*) FROM @meta_table) 
IF @numrows > 0 
    WHILE (@i <= (SELECT MAX(idx) FROM @meta_table)) 
    BEGIN 
        SET @col = (SELECT COLUMN_NAME FROM @meta_table WHERE idx = @i)

        -- do something with @col

        SET @i = @i + 1 
    END

In a first step I’d like to check all columns but something like this does not work

IF (SELECT @col FROM inserted) <> (SELECT @col FROM deleted)
BEGIN
    -- INSERT into logging table ...
END

Additionally that would only check the first row of updates, so I would need to do that for each row in the deleted/inserted table.

Answer 1

Change Data Capture can be a solution to log table DML changes.