Add `dt.prod()` test for a grouped column, `by()` docs adjustment

[samukweku]

• the fix for #3390 has been already pushed as a part of #3388, here we merely add a corresponding test;
• as of #2472, f[:] excludes the groupby columns, in this PR we make the corresponding adjustments to the docs.

Closes #3390

[oleksiyskononenko]

So it looks that we already had a test in test_reduce.py that should caught the bug

def test_prod_grouped():
    DT = dt.Frame(A=[True, False, True, True], B=[None, None, None, 10], C=[2,3,5,0.1])
    RES = DT[:, prod(f[:]), by(f.A)]
    REF = dt.Frame(A=[False, True], B=[1, 10]/dt.int64, C=[3,1.0]/dt.float64)
    frame_integrity_check(RES)
    assert_equals(RES, REF)
    assert str(RES)

However, prod(f[:]) actually calculated products for B and C columns only... even though one may think that f[:] means all the columns, including the grouped one.

[samukweku]

For groupby , f[:] excludes the grouping column - the user would have to explicitly add the grouping column to the j section

[oleksiyskononenko]

Yes, so I am thinking if we ever documented this behavior. We probably need to adjust:

https://datatable.readthedocs.io/en/latest/api/dt/f.html

and

https://datatable.readthedocs.io/en/latest/api/dt/by.html

to mention that f[:] means different things with and without groupby.

[samukweku]

Was handled here : #2472 ... Not sure if it was documented in the groupby docs

[samukweku]

Yea, I will update the docs to reflect that

[samukweku]

If jis a "select-all" slice (i.e.:), then those columns will also be excluded from the list of all columns so that they will be present in the output only once.

The above seems to cover this and can be found here : https://datatable.readthedocs.io/en/latest/api/dt/by.html

[oleksiyskononenko]

It doesn’t refer to f[:], but only to :. At the same time, on the f-documentation page we say f[:] is all columns.

[oleksiyskononenko]

Anyways, don't worry about that, let me just push a minor commit to this PR and we're all set.

[oleksiyskononenko]

@samukweku I made some minor adjustments to this PR, see if it looks good to you.

[samukweku]

Looks great! Thanks @oleksiyskononenko