Yes, using the id as a key is the way to go On Tuesday, November 3, 2015, Seth Axen <[email protected]> wrote:
> Good to know! Thanks so much for the explanation! So for now, it's much > safer to use the atom or conformer id as, for example, a dict key than to > use the atom or conformer itself. > Seth > > On Tue, Nov 3, 2015, 7:06 AM Greg Landrum <[email protected] > <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: > >> Hi Seth, >> >> >> On Tue, Nov 3, 2015 at 7:26 AM, Seth Axen <[email protected] >> <javascript:_e(%7B%7D,'cvml','[email protected]');>> wrote: >> >>> I was wondering why, when I request the conformers (or atoms) from a >>> mol, and then request the owning mol of the conformer or atom, the two mols >>> are not the same instance: >>> >>> mol.GetConformers()[0].GetOwningMol() == mol >>> False >>> >>> I'd expect that they would be equal. Or, is there another way to >>> retrieve the mol instance from an atom or conformer instance? >>> >> >> They actually are the same, it just doesn't look like they are because >> the pointers are different (this is an implementation detail). >> Here's a little demo showing this for atoms : >> >> In [26]: m = Chem.MolFromSmiles('CCC') >> >> In [27]: at = m.GetAtomWithIdx(0) >> >> In [29]: at2 = m.GetAtomWithIdx(0) >> >> >> # they look different: >> >> In [28]: at >> Out[28]: <rdkit.Chem.rdchem.Atom at 0x106031d48> >> >> In [30]: at2 >> Out[30]: <rdkit.Chem.rdchem.Atom at 0x106031528> >> >> In [31]: at==at2 >> Out[31]: False >> >> >> # but are the same: >> >> In [32]: at.SetAtomicNum(7) >> >> In [33]: at2.GetAtomicNum() >> Out[33]: 7 >> >> >> >> It's certainly worth exploring whether or not it's possible to make this >> a bit more transparent and enable either "==" or "is" to work with these >> objects. That investigation is now on the todo list. :-) >> >> -greg >> >>
------------------------------------------------------------------------------
_______________________________________________ Rdkit-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
