We thought about that, but particularly with `GROUP BY ALL` the problem is that we would get different results from SQLite. When a column is not mentioned in the `GROUP BY` clause in SQLite it automatically pushes a `FIRST` aggregate over that column. So for example, the following query:

  SELECT city, COUNT(*)
  FROM customers

In SQLite is transformed into:

  SELECT FIRST(city), COUNT(*)
  FROM customers

In our experience this is not a good default since it is almost never what you want, and hence we did not copy this behavior and instead throw an error in this situation. However, if we were to add an implicit `GROUP BY ALL` our transformed queries would now diverge, i.e. we would transform the above query to:

  SELECT city, COUNT(*)
  FROM customers
  GROUP BY city

Having diverging query results from SQLite on quite basic queries would confuse a lot of newcomers in DuckDB, and potentially cause silent problems when query semantics change when switching databases.

We could definitely add a flag to enable this behavior, however.

I vaguely remember that this is what mysql does?