Why mlforecast.predict does not handle NA values?

[Ehsan86]

To just give a bit context here, I always receive data with a 3-day delay. This means that at time T. I have data only from T-4 back to T-100. Therefore, I do not have access to data for T-1, T-2, and T-3. Given that, I want to train a model based on data from T-100 to T-3 and then test it for time T. How can MLForecast handle this? When I use mlforecast.predict it gives me an error indicating that the data is missing, referring to the missing data from T-1 to T-3. Can anyone help me with this issue as my train data and test are not continuous?

[jmoralez]

Same as #371