Hello!
In my columns i use such calculation DATETRUNC([Group by ], [date])
Where [Group by ] is a parameter that can be day, week, month or year.
And i can't forecast because of this. If i use DATETRUNC('month', [date]) - it works.
But i need opportunity to change my date by parameter, and i want my forecast to be changed with it too.
Is it possible ?
Agree with Adrian that it is a conflict with date granularity.
By default, the forecast is set to Automatic (to the level of date granularity in the view). Separately for the forecasting options dialog box, there are options for changing the date granularity via the dialog options box.
Per the Help Article on Troubleshooting Forecasting, they must match. So, my opinion is that by using a parameter, it causes a conflict between what's in the view versus that of what's in the forecasting options dialog box.
So, you might want to consider Predictive Modeling as a different option, which would likely allow you to change the Model Percentile or in your case Model Quantile() functions with a parameterized date level.
Just know that there are differences between Forecasting (exponential smoothing) and Predictive Modeling (generally linear regression).
Best, Don
