Angles in data file

Hello guys,
    I have a wired problem with angles in the lammps data file.
    For example, if I have a water molecule, then the angles should be
written as
    1 1 1 3 2. (HHO,atomID is 1,2,3.)
    Then I add an oxygen molecule under the water, the angles should be
    1 1 1 2 3
    2 2 4 5.(O2,atomID is 4,5.)
    It worked well.

it may have run, ​but it is incorrect. there is no angle in a diatomic and
you switched the order​ of the atoms for the water molecule.
adding an angle definition by itself doesn't do anything unless you also
change the number of angles in the header. *had* you done that, you would
get the same error as given below.

    But when I have a system with pure oxygen, it always failed with the
message invalid atomID in angles section.
    I read from manual that the syntax for angles is ID type atom1 atom2
    So does it mean that when I write angles in data file, there must be
three atoms in an angle?
    But if it is, why my water and oxygen data worked well surprisingly?
    I know an angle for a two-atom molecule is 180 degree. I tried to
delete the angles of oxygen in data file and it also didn't work.

​this is complete nonsense. an angle potential requires 3 atoms.