Changed features and components in Planning Analytics Engine

ODBC proxy service

Where earlier versions of the Planning Analytics v11 database included an ODBC proxy for increased fault tolerance, Planning Analytics Engine provides an ODBC ‘proxy’ service that is installed close to the data itself. This service is known as the Planning Analytics ODBC Data Connector and it can be installed anywhere as long as the Planning Analytics Engine can access it over the WAN.

Note that the Planning Analytics ODBC Data Connector is a Windows service. It does not run in the containerized environment but rather on a Windows machine with the required ODBC driver set up and access to the data source itself.

For more information, see Using the Data Connector.

Database names are case-sensitive

Database names are case sensitive in Planning Analytics Engine. This is a change from the Planning Analytics v11 database.

The main impact is to Planning Analytics for Microsoft Excel reports and websheets, where the name of the database is referenced in Planning Analytics worksheet functions. Planning Analytics for Microsoft Excel reports must ensure to use the exact case when referencing a Planning Analytics Engine database name.

It's important to note that while Planning Analytics Engine database names are case-sensitive, Planning Analytics Administration does not allow databases that have the same name with different cases. Renaming databases is not possible, you must delete and recreate if you need to rename a database.

TurboIntegrator error log file changes

The error log file mechanism of TurboIntegrator has been updated in Planning Analytics Engine.

In the Planning Analytics v11 database, a TurboIntegrator error log file would be generated only if an error was encountered during process execution. Each process had its own error log file, whether it was an independent process or a process that was asynchronously launched by a parent process.

In Planning Analytics Engine, a process that is launched by either an HTTP request or a chore always generates an error log file, regardless of whether an error occurred during execution. Further, all the child process it launches will aggregate all their errors into the error log file of the main process. The name and format of the error log file is the same as in Planning Analytics v11, with a partial UUID appended to the process name to differentiate the same process executed on the same thread with identical timestamp.

To clearly identify the source of an error during process execution, all the error messages generated by a child process will have the process name appended to the error message:

child_process_1: Error: Failed to create "Dim1". // An error occurred during child_process_1 execution 
Error: Prolog Line (1).... // An error occurred during main process execution The error log file would be empty if no errors occurred during its execution. 

When the user session is destroyed either by api/v1/ActiveSession/tm1.close or because of a timeout, all the empty error log files (which indicate a successful process execution) are cleared with the session.

TurboIntegrator function changes

Additionally, AsciiOutput and TextOutput no longer writes byte order marks (BOMs).