How to combine datasets with different time granularity?

Hi!

I was wondering how could I use Prophet to use two datasets that have yearly and monthly datapoints? I.e. let's assume dataset A has yearly datapoints and dataset B has monthly datapoints. Dataset B contains useful **monthly** features that should be used for predicting variable Y in dataset A. Dataset A also contains useful **yearly** features that help with predicting variable Y in dataset A. If I right join two datasets together by some common key, I will have a lot of **NaNs** because dataset A has only yearly data.

How should I approach this task with Prophet? Should I use some imputation method for that, or does Prophet have some functionality to overcome it? I was also thinking to summarize monthly data into "yearly" representation but that would lead to information loss.

Thanks in advance for help!