We assume that the new SaadaDB is named MYSAADADB

 First create a database in Postgres or make sure it is completely empty if it exists.

• saadaop@host% createdb MYSAADADB
• The name of the postgresql base is not necessary the same as the one of the SaadaDB even if it is quite recommended.

 Run SaadaDB creation tool.

• Go in SAADA_HOME/bin
• Type the command : saadaop@host% newDB The command starts with a Java stack trace and finally open this panel.

New SaadaDB screenshot

 Name of the new database : This name must contain letters, mumerics and ’_’ only. Take care that no other web application on your Tomcat instance has the same name otherwise it will be scratched.

 DBMS database name : This is the name of the PSQL base name. We advice you to keep the SaadaDB name if possible.

 DBMS user name : The PSQL user name is used to connect to the database. Usually it is the saadaop user name, which is also the person who is creating the database.

 DBMS user password : There is no password by default on Postgres. If you want to set up one, mention it here.

 DBMS server name or IP : If you have installed Postgres on your machine, keep locahost.

 DBMS port number : If the database server runs on another machine, type here the TCP port used by the connection (5432 is the default port for Postgres)

 Web server host name : Filling properly this field is very important for running the Web interface. Never let the default localhost.

• If you get directly access to your SaadaDB on your Tomcat instance, put http://your_host:8080/SAADA_DB_NAME. We assume here that your Tomcat used the port 8080 and that your SaadaDB is named SAADA_DB_NAME.
• If you get access to your Tomcat instance through a proxy, type the proxy address.

 Select a database Driver : Only Sybase and Postgres are available yet. Note that Sybase has become obsolete.

 Coordinate System : Select here the coordinate system used at collection level. Loaded product can have coordinates expressed in various systems. These native coordinates can thus not be used to make selection on multiple data classes. For this reason, Saada converts native coordinates into collection system. These computed positions, stored at collection level, are used by the query engine.

 Optional Product Units : This (badly named) form only concerns spectra. As for coordinates, you can choose a spectral coordinate system which will be used for all loaded spectra. That allows the query engine to make spectra selections over multiple classes with constraints on the spectral range. The spectral coordinate conversion can fail due to the huge range of possible definitions. In this case, spectra are nevertheless loaded but spectral coordinate columns are not populated at collection level.

 Select a Saada Root Dierctory : This file chooser selects the directory where the Saada stuff (code and resources but not the data) will be installed (see working area setup).

 Select a Saada Repository Dierctory:This file chooser selects the directory where data files loaded in your SaadaDB will be copied (see working area setup).

 Add Attributes : This button opens a popup window setting up new atributes at collection level (see Add new Saada Attributes).

 Console : Nothing appears on the console in this context, look at your termimal indead.

 Create Saada Database : CLick on this button to create the database.

• This operation consists in copying a directory tree into SAADA_DB_HOME, generating and compiling some codes and creating some directories. There is no access to the Postgres database.
• The SaadaDB created is logged in the file SAADA_DB_HOME/config/list_saadadb.xml. If you want to create your SaadaDB (with the same name) again, check first that the entry matching your SaadaDB name has been removed from that file, then you can remove it by hand if needed.