Since the launch of Planning Analytics few years back, IBM has been recommending its users to move to Planning Analytics for Excel (PAX) from TM1 Perspective and TM1 Web. As every day new users migrate to adopt PAX, it’s prudent that I share my experiences.
This blog will be part of a series where I would try to highlight and make users aware of different aspects of this migration. This one specifically details a bug I encountered during one of the projects in which our Clients was using PAX and steps taken to mitigate the issue.
What was the problem:
Scenario: when a Planning Analytics User triggers a process from Navigation Pane within PAX and uses “Edit parameters” option to enter value for a numeric parameter and clicks save to runs the process.
Issue: when done this way, the process won’t complete and fail. However, instead if this was run using other tools like Architect, Perspective or TM1 Web, the process would complete successfully.
For example, let’s assume a process, cub.price.load.data takes a number value as input to load data. User clicks on Edit Parameter to enter value and saves it to run. The process fails. Refer screenshots attached.
What’s causing this:
During our analysis, it was found that while using PAX, when users click on Edit parameter,enter value against the numeric parameter and save it, in the backend the numeric parameter was getting converted into a String parameter thereby modifying the TI process.
As the TI was designed and developed to handle a numeric variable and not a string, a change in type of the variable from Numeric to String was causing the failure. Refer screenshots below.
What’s the fix?
Section below illustrates how we mitigated & remediated this bug.
For all TI’s using numeric parameter.
- List down all TI’s using numeric type in Parameter.
- Convert the “Type” of these parameters to String and rename the parameter to identify itself as string variable (best practice). In the earlier example, I called it pValue while holding numeric and psValue for String.
- Next, within the TI in Prolog, add extra code to convert the value within this parameter back to same old numeric variable. Example, pValue =Numbr(psValue);
- This should fix the issue.
Note that while there are many different ways to handle this issue, it best suited our purpose and the project. Especially considering the time and effort it would require to modify all effected processes.
Planning Analytics for Excel : Versions effected
Latest available version (as of 22ndOctober 2019) is 2.0.46 released on 13thSeptember 2019. Before publishing this blog, we spent good time in testing this bug on all available PAX versions. It exists in all Planning Analytics for Excel versions till 2.0.46.
Permanent fix by IBM:
This has been highlighted to IBM and explained the severity of this issue. We believe this will be fixed in next version of Planning Analytics for Excel release. As per IBM (refer image below), seems fix is part of the upcoming version 2.0.47.
You may also like reading “ Predictive & Prescriptive-Analytics ” , “ Business-intelligence vs Business-Analytics ” ,“ What is IBM Planning Analytics Local ” , “IBM TM1 10.2 vs IBM Planning Analytics”, “Little known TM1 Feature - Ad hoc Consolidations”, “IBM PA Workspace Installation & Benefits for Windows 2016”.