Default output column name for count distinct is inconsistent with the rest

- `sum("a", 0, 1)` → column is named `a_sum`
- `average(a", 0, 1)` → column is named `a_average`
- `count_distinct("a")` → column is named… you guessed it, `count_distinct(a)` 🙃

Honestly though, the count_distinct syntax is much nicer (and more consistent with PySpark). @Maegereg @tmager what would you think of using the parenthesis syntax by default everywhere?