Also keep in mind that having a NULL value vs not having any data at all are two different things. A NULL value would be like having a row for Jan 1 2022, but the column value for Sales is empty. This is different than not having any rows for Jan 1 2022 at all. In the former scenario, you can replace the NULL with a 0 as some have pointed out on this thread (with an ifelse kind of calculation). In the latter scenario where there are no rows at all in the data, there is nothing to insersert a substitute into.
QuickSight (and most BI tools) being read-only, we cannot insert new rows of data (can only make new columns). I think this is where your question about doing UNIONS in the data prep is probably coming from? You are right that would be a solution to the latter scenario (to insert a row with a null or 0 value), but as other mentioned the only want to to a UNION is by using Custom SQL. Yes, we are looking at adding UNIONS into the data prep UI as well, but unfortunately I cannot comment on dates in this forum 