Improve Error message when calling aggregation method on dataframe without any columns

[samster25]

Currently when run an aggregation function on a dataframe without input columns:

df = daft.from_pydict({"x": [1,2,3]})
df.count()

we get the following error message, which is not very helpful.

AssertionError: no columns were passed in

We should instead:

1. run the aggregation on all columns
2. give a better error that give an example
3. tell the user to look for df.count_rows() instead.

[colin-ho]

I can take this!