I have noticed that when Indigo nodes read in an SDF into Indigo format, and then when its converted back to SDF format, all the original atom numbering has been scrambled compared to the original SDF. This is really problematic when you want to specifically deal with individual atoms in the molecule.
CDK and RDKit nodes preserve the atom numbering, and would be ideal if Indigo could do the same.
You can try the latest nightly build where this issue is resolved. But we had to change our internal format and all the intermediate data (with Indigo objects) should be reevaluated. Without it you will get “indigoUnserialize(): format not recognized” exception.