As part of dealing with a locking problem I am fine tuning a stored procedure
that updates a table. The application updates a row by changing every single column except for the primary key, whether 1 or all of the columns have
been modified. Although easier to code, this strikes me as taking a sledge hammer to a nut.
Could anyone tell me if there is any benefit in attempting to break this down. That is, code the stored procedure so that only those columns being changed are modified. I am thinking this might reduce dramatically the overhead of writing to the transaction log and making the changes to the actual row.If the benefit is non existant (or insignificant) because of the way Sql Server updates a row it will obviously be a waste of time to generate