Having NULL values for cd_smiles is not a problem in itself:
These columns store the standardized structure in a compact format, allowing efficient caching and hence fast structure searching. (If this representation of the structure is larger than the maximum length of the column or cannot be represented for any other reason, then
NULL is stored and the
cd_structure field is used during the search.)
So only those rows are affected where there was actually an exception.
We could reproduce the exception with the attached structures (it is enough to try to open in MarvinSketch).
At the moment it seems to be a possible bug in our molfile import, or some problem in the molfile which was previously overlooked by our code.
(5.2.x versions seems to read it, 5.3.x versions not)
How can I fix these rows? Is there a way to regenerate the cd_smiles column using the cd_structure column's data?
We are always using the cd_structure column during regeneration (all other calculated columns are updated, but cd_striucture never changes).
One way is to downgrade to a previous version, but that usually requires manual intervention in the database.
Please let us know if you need instructions.
We are just before issuing a patch release, so if you can wait a couple of days we may quickly fix this and include in this patch.
In general pre-regeneration (or pre-calculation) is an excellent way to avoid any surprises and also reduce offline time due to recalculation:
We'll get back to you with our findings about these structures ASAP, we are sorry for the inconvenience.