[Rdkit-discuss] nitro-compounds from smarts

[Rafal R. <rmr...@gm...>]

hi all,

Is it possible to read smarts with nitro group in form 'N(=O)(=O)'?

Here is the context of my question:
The example below shows that there is some problem with nitro group and
MolFromSmarts: 

>>> m1=Chem.MolFromSmiles('c1ccccc1[N+](=O)[O-]')
>>> m2=Chem.MolFromSmiles('c1ccccc1N(=O)(=O)')
>>> m11=Chem.MolFromSmarts('c1ccccc1[N+](=O)[O-]')
>>> m22=Chem.MolFromSmarts('c1ccccc1N(=O)(=O)')
>>> Chem.MolToSmiles(m1)
'O=[N+]([O-])c1ccccc1'
>>> Chem.MolToSmiles(m2)
'O=[N+]([O-])c1ccccc1'
>>> Chem.MolToSmiles(m22)
'O=N(=O)c1ccccc1'
>>> Chem.MolToSmiles(m11)
'O=N(O)c1ccccc1'

Function MolFromSmiles accept both form ('[N+](=O)[O-]' and 'N(=O)
(=O)') and return the same canonical smiles. This make sense to me.
The behaviour of MolFromSmarts is strange, nitro in form '[N+](=O)
[O-]' is correctly detected but return different canonical smiles (why?
). However 'N(=O)(=O)' form is interpreted incorectly.
Is this bug or feature? 
I have a lot of smarts which can have nitro group written either
as 'N(=O)(=O)' or '[N+](=O)[O-]'. What should I do to correctly read
them all? Replace all 'N(=O)(=O)' string (and all possible variations)
to ionic form?

Best,

Rafał