Chapter 3 Connecting to the repository

Initializing the repository

When you connect for the first time to a database that will store the repository, you initialize the repository and create the tables. When the tables are created, they are owned by the user logged to the database. Initialization messages appear in the Generation tab in the Output window.

Every repository user shares the same database user login to access database tables.

First repository connection
Whatever the repository user login name you have specified in the Connect dialog box, the first time you connect to the database you are connected as ADMIN in order to be able to create the repository tables.

Repository installation script

Initialization and upgrade statements appear first in the Repository Installation Script window. You can use this window to browse the statements.

If you need to modify the statements (add physical options for example) you have to copy the text, and paste it in a script interpreter. The Separator dropdown listbox lets you select the separator that corresponds to your interpreter. When you change separator using the Separator dropdown listbox, the script automatically displays this change. You then have to click Cancel in the Repository Installation Script window, then modify and execute the statements with your script interpreter.

WARNING!

Read-only window
The Repository Installation Script window is read-only. You cannot edit any statement.

Repository installation failure

Repository installation may fail for the following reasons:

Network problem.
Insufficient disk space or cache problem.
Permission problem.
Conflicts with existing objects.
...

In case of failure, a warning message appears with the following buttons:

Ignore: allows you to skip the statement that causes an installation problem and resume installation, the repository is partially installed.
Ignore All: allows you to skip all statements that cause an installation problem and resume installation, the repository is partially installed.
Abort: stops the installation process and drops the tables already created (see Dropping the repository tables).

It is recommended to drop repository tables and restart installation in case of network problem, insufficient disk space, or permission problem.

WARNING!

Caution
Do not drop repository tables in case of conflicts with existing objects as drop process may remove tables with identical names from other applications in the database. Also, avoid using the Abort option when conflicts with existing objects occurs as Abort option also drops repository tables and tables with identical names.

In case of conflicts with existing objects, you should:

Identify conflicting tables during script execution. You should run the initialization script and use the Ignore button to identify all the conflicting tables.
Once you have the list of the conflicting tables, you should remove them from the repository tables drop script. To do so, you should run the Drop Repository Tables feature, copy the script text and paste it in a script interpreter, and remove the conflicting tables.
You can then run the repository tables drop script. This will remove all the repository tables except the conflicting tables.
You can now remove conflicting tables in the existing database and restart repository installation.

or

Install the repository in another database that does not contain conflicting tables.

Repository database size

Model versions are usually stored as a set of differences from the previous version. They share the unmodified object versions with the previous version. This does not apply to the first versions (v1 or v1.1), or the baselined versions, that are both stored as a set of fully independent objects. As a rule of thumb, such an independent version takes in the database twice the size of the corresponding model stored in an XML file.