When changing the primary key of a table you change the very
definition of a the data stored in that table. Typically, this should
never happen if you have a good data model/design that represents
logical business entities in your organization. But in the real
(non-academic) world, this could be required. Using the primary key of
a table, you should be able to uniquely identify a particular instance
of a record in this entity. Changing the key means that you have
changed your definition of the entity, therefore you really do need to
go into every program that accesses this table and change it, since
you have altered the very definition of the data within this table.
In a nutshell, I recommend not doing anything fancy, and going into
the code to start hashing through it. The only shortcut around this
(especially if the data is maintained via lots of programs) is to
create a before insert trigger (if the new column is null, then you
can insert a default maybe)? You will also have to play with before
update triggers as well. Also, you will have to change any of the
update programs, etc... All these triggers opens up a whole new can
of worms. Good Luck...Rich (http://www.sqlquery.com)
Received on Thu Jul 17 2003 - 12:10:12 CDT
