IBM SPSS Modeler Server and the data access pack

To configure ODBC to start with IBM SPSS Modeler Server

1. Stop the IBM SPSS Modeler Server host if it is running.
2. Download the relevant compressed TAR archive for the platform on which you have IBM SPSS Modeler Server installed. Make sure that you download the correct drivers for your installed version of IBM SPSS Modeler Server. Copy the file to the location where you want to install the ODBC drivers (for example, /usr/spss/odbc).
3. Extract the TAR archive file using tar -xvof.
4. Run the setodbcpath.sh script that is extracted from the archive.
5. Edit the script odbc.sh to add the definition of ODBCINI to the bottom of this script and export it, for example:

   ODBCINI=/usr/spss/odbc/odbc.ini; export ODBCINI

   ODBCINI must point to the full pathname of the odbc.ini file that you want IBM SPSS Modeler to read to get a list of the ODBC datasources that you define (a default odbc.ini is installed with the drivers).

6. Save odbc.sh.
7. (64-bit IBM SPSS Modeler Server installations only; for other installations, continue from the next step) Define and export LD_LIBRARY_PATH_64 in odbc.sh:

   if [ "$LD_LIBRARY_PATH_64" = "" ]; then
             LD_LIBRARY_PATH_64=<library_path>
   else
             LD_LIBRARY_PATH_64=<library_path>:$LD_LIBRARY_PATH_64
   fi
   export LD_LIBRARY_PATH_64
   

   where library_path is the same as for the LD_LIBRARY_PATH definition already in the script that has been initialized with your installation path (for example, /usr/spss/odbc/lib). The easiest way to do this is to copy the if and export statements for LD_LIBRARY_PATH in your odbc.sh file, append them to the end of the file, and then replace the "LD_LIBRARY_PATH" strings in the newly appended if and export statements with "LD_LIBRARY_PATH_64".

   As an example, your final odbc.sh file on a 64-bit IBM SPSS Modeler Server installation might look like this:

   if [ "$LD_LIBRARY_PATH" = "" ]; then
             LD_LIBRARY_PATH=/usr/spss/odbc/lib
   else
             LD_LIBRARY_PATH=/usr/spss/odbc/lib:$LD_LIBRARY_PATH
   fi
   export LD_LIBRARY_PATH 
   if [ "$LD_LIBRARY_PATH_64" = "" ]; then
             LD_LIBRARY_PATH_64=/usr/spss/odbc/lib
   else
             LD_LIBRARY_PATH_64=/usr/spss/odbc/lib:$LD_LIBRARY_PATH_64
   fi
   export LD_LIBRARY_PATH_64 
   ODBCINI=/usr/spss/odbc/odbc.ini; export ODBCINI 

   Remember to export LD_LIBRARY_PATH_64, as well as defining it with the if loop.

8. Edit the odbc.ini file that you defined earlier using $ODBCINI. Define the data source names that you require (these depend on the database that you are accessing).
9. Save the odbc.ini file.
10. Configure IBM SPSS Modeler Server to use these drivers. To do so, edit modelersrv.sh and add the following line immediately below the line that defines SCLEMDNAME:

    . <odbc.sh_path>

    where odbc.sh_path is the full path to the odbc.sh file that you edited near the beginning of this procedure, for example:

    . /usr/spss/odbc/odbc.sh

    Note: The syntax is important here; be sure to leave a space between the first period and the path to the file.
11. Save modelersrv.sh.

To test the connection

1. Restart IBM SPSS Modeler Server.
2. Connect to IBM SPSS Modeler Server from a client.
3. On the client, add a Database source node to the canvas.
4. Open the node and verify that you can see the data source names that you defined in the odbc.ini file earlier in the configuration procedure.

If you do not see what you expect here, or you get errors when you try to connect to a data source that you have defined, follow the Troubleshooting procedure. See the topic Troubleshooting ODBC configuration for more information.

To configure ODBC to start with IBM SPSS Modeler Solution Publisher Runtime

When you can successfully connect to the database from IBM SPSS Modeler Server, you can configure an IBM SPSS Modeler Solution Publisher Runtime installation on the same server by referencing the same odbc.sh script from the startup script of IBM SPSS Modeler Solution Publisher Runtime.

1. Edit the modelerrun script in IBM SPSS Modeler Solution Publisher Runtime to add the following line immediately above the last line of the script:

   . <odbc.sh_path>

   where odbc.sh_path is the full path to the odbc.sh file that you edited near the beginning of this procedure, for example:

   . /usr/spss/odbc/odbc.sh

   Note: The syntax is important here. Be sure to leave a space between the first period and the path to the file.
2. Save the modelerrun script file.
3. By default, the DataDirect Driver Manager is not configured for IBM SPSS Modeler Solution Publisher Runtime to use ODBC on UNIX systems. To configure UNIX to load the DataDirect Driver Manager, enter the following commands (where sp_install_dir is the installation directory of Solution Publisher Runtime):

   cd sp_install_dir 
   rm -f libspssodbc.so
   ln -s libspssodbc_datadirect.so libspssodbc.so

To configure ODBC to start with IBM SPSS Modeler Batch

No configuration of the IBM SPSS Modeler Batch script is necessary for ODBC. This is because you connect to IBM SPSS Modeler Server from IBM SPSS Modeler Batch in order to run streams. Ensure that the IBM SPSS Modeler Server ODBC configuration has been made and is working correctly, as described earlier in this section.

To add or edit a data source name

1. Edit the odbc.ini file to include the new or changed name.
2. Test the connection as described earlier in this section.

When the connection with IBM SPSS Modeler Server is working correctly, the new or changed data source should also work correctly with IBM SPSS Modeler Solution Publisher Runtime and IBM SPSS Modeler Batch.

SQL Server support with the Data Access Pack driver

The ODBC configuration for SQL Server must have the Enable Quoted Identifiers ODBC connection attribute set to Yes (the default for this driver is No). On UNIX this attribute is configured in the system information file (odbc.ini) using the QuotedId option.