Forum OpenACS Q&A: loading new data model

One file must be named "yourpackagekey-create.sql" in order for the APM to react. Similarly, if you have drop scripts, one of them must be "yourpackagekey-drop.sql". The create file usually doesn't contain much of interest. What it usually does contain is stuff like:

/i yourpackagekey-tables-create.sql
/i yourpackagekey-packages-create.sql

on PostgreSQL ... or

@ yourpackagekey-tables-create.sql
@ yourpackagekey-packages-create.sql

on Oracle.

These lines will pull in the scripts that actually do something - but I'm sure you knew that part already.

After you have renamed the file, you should watch/reload the files of your package or restart the server.

BTW, Sometimes you see something like "@@ blah.sql", i.e., two @s instead of just one ... Can someone tell me what the difference is? I tried to Google but I couldn't find an answer.

/Ola

Just more info ... if you've added a .sql file and intend to drop and recreate your package instance (or entire OpenACS install) because you're in development, simply adding it to your <foo>-create.sql file is fine, as mentioned above.

If however recreating the package is inconvenient (it has live data, etc) you can either load the SQL file by hand or bump your package version number in the .info file, and add a file named "upgrade/upgrade-<old version>-<new version>.sql" under your sql/postgresql/ directory.  You'd then visit the APM, click on "install packages" and presto! you'll see your package offered labelled "upgrade" rather than "install".  This upgrade file would itself then reference your new .sql file.