One of the least-often-mentioned aspects of Agile Development (IMO) are the challenges involved in database change management, described in Martin Fowler's
Evolutionary Database Design.
Martin states "One of the most valuable pieces of automation is a simple set of scripts for common database tasks" - but I'm not so sure that there is a simple way to accomplish all that in a foolproof manner.
It's definitely do-able in a team of disciplined developers with good SQL knowledge, experience and dedication to writing DDL, DML and understanding when they're making a destructive change. But what about teams of different skill levels (particularly in SQL), very large databases where mulitple delta-changes, when applied, will take HOURS to run (ie. longer than a maintenance window), etc...?
Not that I have a solution right now - but why isn't it a more visible "problem" - what is everyone doing to solve this problem - manual processes?????
[EDIT 9-May-05] Peter Hancock's Continuous Database Integration article describes a novel, NAnt-based approach; but it still requires a lot of 'thought' to ensure it always works...
[EDIT 15-May-05] Unit Testing the Data Access Layer covers the 'related' issue of how to test the database...