#db-ratchet User's Guide
##Schema Management Schema authoring support in db-ratchet is supported by use of a SchemaDifferenceEngine. A SchemaDifferenceEngine handles comparing databases and file store representations of databases to produce version upgrade scripts. Currently db-ratchet has a single SchemaDifferenceEngine implementation based on Redgate SQlCompare. There are near term plans to add another implementation based on XML representation of schema objects that will not require commercial 3rd party software.
The following diagram illustrates how schema management operations interact with the the database repository.
Builds of database schema changes are handled by Flyway DB.
##Data Management db-ratchet has 2 operations to support data migrations: capture and migrate. These operations manage data packages in an all or nothing fashion, meaning there is no support for doing partial or row level migration of data. A data package is defined as one or more tables that need to be migrated as an atomic unit.
This form of data migration support is intended for control data whose life cycle is tied to that of application development and whose migrations to a database happen in connection with a deployment of software and schema changes. Data packages are stored in the data directory of a database repository and are configured in /data/data-packages.xml. Please see the comments in the template data-packages.xml in your database repository for instructions on configuring data packages.
Data is stored in the data/packages directory and is based on the DB Unit XML format.
Data migrations leverage temp tables and SQL MERGE statements to minimize the work that is done inside of the boundaries of a database transaction. With this in mind we should be able to perform migrations for modest datasets without the need to shut down the system reading from the database. The migration of a database package will succeed or fail atomically, so the possibility of leaving behind an inconsistent dataset is eliminated.
The choice to use SQL MERGE statements was based on this tool initially being developed for management of SQL Server databases only. Future plans for this project include developing an alternate data migration implementation for other database platforms.
Interested in working with Markdown? Get started with this Chrome extension for viewing .md files: Markdown Preview