schema versioning... best practices... ideas

Subject:schema versioning... best practices... ideas
Date:Sun, 3 Jan 2010 11:00:12 -0800 (PST)
The normal way I've seen database schema versioning is to create a
database version table and use an integer as the schema version.

I've found a few instances where applications use the product version
also as the schema version.

I like the latter because a specific version of my application must
only be used with a specific version of our schema, and so it keeps
things simple: I just have to change one version in my application,
and then write my upgrade logic based on that for both the software
and the database schema. Each build changes the product version
(build number is part of the product version), and so for any software
change there is always a different product version, and so maintaining
two separate versions seems like unnecessary work.

Are there any documents, web sites..., that explain the pros and cons
for the different ways of doing schema versioning?

Even when I use the product version, that can be translated into an
integer, and so it is still a single number, but just formatted in
n.n.n.n format.



Other posts:
EISWT-10 Call for papers
Query to obtain (A AND (NOT B))?
[ANN] DW Data Concepts announces fmWebMaze.com
Indici
Keyword search in spatial database
• schema versioning... best practices... ideas
FirstSQL/J 3.20 Embedded Mobile RDBMS for JME
Call for Repository Content about Combinatorial Optimization Problems & meta-heuristics
IS 2010 (2nd Call) submissions until 8 January 2010
Finding tables in Open Office Base
Call for Papers: The 2010 International Conference on Information and Knowledge Engineering (IKE'10), USA, July 2010

generated at 15:43:31