Computed columns within groupby may appear shuffled

[st-pasha]

>>> from datatable import *
>>> DT = Frame(A=[1, 2, 1, 3, 2, 2, 2, 1, 3, 1], B=range(10))
>>> # This is correct
>>> DT[:, [f.A + f.B], by(f.A)]
      A  C0
---  --  --
 0    1   1
 1    1   3
 2    1   8
 3    1  10
 4    2   3
 5    2   6
 6    2   7
 7    2   8
 8    3   6
 9    3  11

[10 rows x 2 columns]

>>> # This is correct too
>>> DT[:, [f.A + f.B, f.B], by(f.A)]
      A  C0  C1
---  --  --  --
 0    1   1   0
 1    1   3   2
 2    1   8   7
 3    1  10   9
 4    2   3   1
 5    2   6   4
 6    2   7   5
 7    2   8   6
 8    3   6   3
 9    3  11   8

[10 rows x 3 columns]

>>> # But this isn't
>>> DT[:, [f.B, f.A + f.B], by(f.A)]
      A  C0  C1
---  --  --  --
 0    1   0   1
 1    1   2   8
 2    1   7   8
 3    1   9  11
 4    2   1   3
 5    2   4   3
 6    2   5   6
 7    2   6   7
 8    3   3  10
 9    3   8   6

[10 rows x 3 columns]