GenDBWiki/AdministratorDocumentation/UpdatingFromOlderVersions

Updating from older GenDB versions

Below you can find some guidelines for migrating older GenDB projects to the latest version.

Updating from GenDB-1

It is of course easy to export a genome annotation from a GenDB-1 project to an EMBL file, but since an EMBL file will only contain the latest annotation you will loose the history of all annotations. Therefore, we have provided a simple Perl script that allows you to migrate the most important data (contigs with all CDS and also tRNAs and their annotations) from a GenDB-1 project into a new GenDB-2 project database. Since the concept for storing tool results was completely redesigned for the GenDB-2.x series, all `facts` (now called `Observations`) stored within a GenDB-1 database will not be moved to the corresponding 2.x project. Instead you will have to recompute all bioinformatics tool results after migrating all project data. In order to avoid inconsistencies during the migration process we advice you to set all user accesses to a read-only status for the GenDB-1 projects so that the old version can't be modified any more during and after running the migration script. If you require a more advanced migration procedure (e.g. for more types of regions) you will have to modify the script and adapt it for your own purposes.

• set DBI ENV for MySQL user name and password, set project name in script
• configure list of all annotators whose annotations should be moved to new project

Of course we can not guarantee a complete and perfect migration of your GenDB-1 data to a new GenDB-2 project, so have your backup prepared and check the newly imported data carefully. It is probably also a good idea to keep the old GenDB-1 projects running for a while in read-only mode so that users can adapt to the new system and check if the migration was successful.

Updating from GenDB-2.0

For migrating data from a GenDB-2.0 project to GenDB-2.2 and also for all future updates we have implemented a more general solution that works with all O2DBI applications. Therefore, we are using generic XML export and import modules that are created automatically based on the O2DBI database schema description. An XML file thus created by exporting e.g. the data from a GenDB-2.0 project is then transformed into the corresponding XML schema of an upgraded version by applying an XSLT transformation. The resulting XML file can then be imported easily into a new project by using the generic import module. This three step process can be performed by executing the following scripts (please take a look at the complete usage message of each script for further details and more options):

1. `export_XML.pl`
2. `transform_GenDB-2.0_to_2.2.pl`
3. `import_XML.pl`

In order to reduce the size of the exported XML and for a faster migration of a projects' data we advise you to delete all observations that are not used as supporting evidence in any annotations. This can be done by running the script `delete_unused_observations.pl`. After migrating the data and importing everything into the upgraded project, all observations have to be recomputed of course (anyway recomputing observations should be done from time to time in order to stay current with the content of the sequence databases).

Of course we can not guarantee a complete and perfect migration of your old GenDB data to a new GenDB project, so have your backup prepared and check the newly imported data carefully. It is probably also a good idea to keep the old GenDB projects running for a while in read-only mode so that users can adapt to the new system and check if the migration was successful.