Imputing null values in time series data

Hi all, I am a novice to PowerBI (and also to forecasting, but not to statistics). I have a data of sales transactions for each week over a whole year for various brands. However, some weeks the sales transactions are Zero. Those (weekly) records when Sales is zero are not a part of the dataset, only non-zero Sales weeks are.

I need to know how PowerBI transforms the data when it is asked to create a simple Forecast Line or Shaded zone, especially considering that those weeks when Sales = 0 are not even in the dataset.

I acquired the forecast lines on both versions of the data -- one, where zero-sales weeks are specifically included, and one where only non-zero sales are reported. The forecast is different in each case.

Shouldn't it be an obvious default step in any forecast module to impute the Zero-Sales weeks with 0s and then plot the line chart? I do not see such an option anywhere. I realize there can be other assumptions made by a forecasting module such as using averages for the missing weeks, and so on, but at the least the option to choose such imputation should be available.

Many thanks,