Structure Checker and broken SDF files

User 677b9c22ff

20-08-2013 02:30:26

Hi,


the Structure Checker v6 does not catch errors from certain broken SD files.


One solution could be to catch the errors that are thrown by MVIEW and use the error output as fix.


 


FILE: error-broken-SDF.sdf


Example: "Invalid counts line" output in mview (its not thrown all the time)


Mview 6 reports 4 molecules, but Structure Checker  v6 thinks  there is only one molecule in the file.


 


Other errors include, that should be fixed by Structure Checker:


Example: chemaxon.formats.MolFormatException: Invalid header block.


Basically any MolFormatException should be fixed.


 


The broken SD files are very common if assembled from multiple sources,


so I think that is an highly probable feature that should be included into the


standard settings. That extends not only to broken headers, but basically


all "crude errors" such as wrong element names, broken SD attachments etc.


 


Other example data can be downloaded here (public)


http://www.smpdb.ca/public/system/current/metabolites.sdf


 


Also if one activates all options in structure checker v6, using the configure checker,


the substructure checker still makes problems because its not marked as requesting input, even if set


ignore one can not proceed with "NEXT". Quite annoying if one has to go throuh all options manually.


 


One other solution I came up with when structure checker fails is to use Instant-Jchem,


because it would catch and ignore such crude errors, maybe similar functionalities could be


implemented into the Structure Checker, that would be quite helpful.


 


Cheers


Tobias

ChemAxon e08c317633

22-08-2013 18:25:41

Hi Tobias,


Structure Checker can report and fix errors in chemical structures. Molecule import from files/streams is handled by MolImporter, Structure Checker can only ignore import errors currently. 


Thanks for the suggestions, we will add import error reporting to Structure Checker application, and we will consider fixing import errors (not just ignoring them) in Structure Checker as well.


Zsolt