Execution Command
For processing engine technology types, you must write an execution command file.
The execution command invokes the external processing system. It must be contained in a file of a type appropriate for the operating system; a shell script for UNIX or a command file for Windows.
In most cases, the execution command must provide a mechanism to pass the service instance to the external processing engine which can then use the service instance's connection credentials to connect back to the Oracle Life Sciences Data Hub database.
To help ensure the proper storage and display of non-English character data in Oracle LSH, the processing engine should use UTF8 character encoding. If you can enforce this in the execution command, do so.
As with the shipped adapters, at each Oracle Life Sciences Data Hub installation the user must move the execution command file into the DP Server directory created by the user, or a subdirectory of it, and enter the actual path of the file in the service location definition in the Oracle LSH user interface (see the Oracle Life Sciences Data Hub System Administrator's Guide). The system picks up the location from the service location definition when a user submits a job.
If you are creating this adapter for use in other companies or locations, you must include instructions for system administrators similar to those in the Oracle Life Sciences Data Hub Installation Guide to move the file and edit it to add whatever information your adapter may require about the location; for example:
- the Oracle SID
- the location of the technology server
- the location of Oracle setup script
coraenv
- a variable holding the path to any external command invoked inside the execution command; for example, if the execution command invokes a Java executable, the path to the Java executable must be set correctly in the path variable.
Parent topic: Planning for Object Execution