A script based load table must have a Host Script defined. During the load process this host script is executed and the results returned.
During the 'drag and drop' creation of a load table from a UNIX or Windows file a script can be generated by selecting one of the 'Script based' load options. This script can then be edited to fully meet any requirements.
If the UNIX Scheduler is not being used and a UNIX script is generated and executed interactively, the following two variables will be setup as part of the environment.
- DSS_USER=username
- DSS_PWD=password
Where 'username' and 'password' are the entries for the data warehouse database.
There are a number of conventions that must be followed if these host scripts are to be used by the WhereScape scheduler.
- The first line of data in standard out must contain the resultant status of the script. Valid values are '1' to indicate success, '-1' to indicate a Warning condition occurred but the result is considered a success, '-2' to indicate a handled Error occurred and subsequent dependent tasks should be held, -3 to indicate an unhandled Failure and that subsequent dependent tasks should be held.
- The second line of data in standard out must contain a resultant message of no more than 256 characters.
- Any subsequent lines in standard out are considered informational and are recorded in the audit trail. The normal practice is to place a minimum of information in the audit trail. All bulk information should be output to standard error
- Any data output to standard error will be written to the error/detail log. Both the audit log and detail log can be viewed from the WhereScape RED tool under the Scheduler window.
- When doing Script based loads, it is easy to use the Regenerate button beside the Script Name field to regenerate the scripts. You can then click the Edit button to view and edit the generated script, as per your requirements.
Script-based loads on Windows support both DOS Batch and PowerShell scripts.