How do you version your database schema? [closed]

Answer 1

If you are still looking for options : have a look at neXtep designer. It is a free GPL database development environment based on the concepts of version control. In the environment you always work with versioned entities and can focus on the data model development. Once a release is done, the SQL generation engine plugged on the version control system can generate any delta you need between 2 versions, and will offer you some delivery mechanism if you need.

Among other things, you can synchronize and reverse synchronize your database during developments, create data model diagrams, query your database using integrated SQL clients, etc.

Have a look at the wiki for more information : http://www.nextep-softwares.com/wiki

It currently supports Oracle, MySql and PostgreSql and is in java so the product runs on windows, linux and mac.

Answer 2

For MySQL

When I land on a new DB:

Firstly, I check structure:

mysqldump --no-data --skip-comments --skip-extended-insert -h __DB_HOSTNAME__ -u __DB_USERNAME__ -p __DB1_NAME__ | sed 's/ AUTO_INCREMENT=[0-9]*//g' > FILENAME_1.sql
mysqldump --no-data --skip-comments --skip-extended-insert -h __DB_HOSTNAME__ -u __DB_USERNAME__ -p __DB2_NAME__ | sed 's/ AUTO_INCREMENT=[0-9]*//g' > FILENAME_2.sql
diff FILENAME_1.sql FILENAME_2.sql > DIFF_FILENAME.txt
cat DIFF_FILENAME.txt | less

Thanks to stackoverflow users I could write this quick script to find structure differences.

src : https://stackoverflow.com/a/8718572/4457531 & https://stackoverflow.com/a/26328331/4457531

In a second step, I check datas, table by table with mysqldiff. It's a bit archaic but a php loop based on information_schema datas make job surely

For versioning, I use the same way but I format a SQL update script (to upgrade or rollback) with diff results and I use version number convention (with several modifications the version number look like an ip address).

initial version : 1.0.0
                  ^ ^ ^
                  | | |
structure change: - | |
datas added: -------- |
datas updated: --------

Answer 3

http://bitbucket.org/idler/mmp - schema versioning tool for mysql, writed in PHP

Answer 4

I am using strict versioning of the database schema (tracked in a separate table). Scripts are stored in version control, but they all verify current schema version before making any change.

Here is the full implementation for SQL Server (the same solution could be developed for MySQL if needed): How to Maintain SQL Server Database Schema Version

Answer 5

After long investigation, I figured out that there are some 3rd party tools or Visual Studio project types that does not satisfy me, or just blogs about the theory but no implementation. So I implemented a working system, which is used almost a year, and explained here:

http://nalgorithm.com/2015/11/09/database-versioning-part-1/

depending on the interest, will continue writing more.

Answer 6

You didn't mention which RDBMS you're using, but if it's MS SQL Server, Red-Gate's SQL Compare has been indispensable to us in creating deltas between object creation scripts.