Strange search behavior

Structures A, B and C are in the database.

 

Structure A: COC(=O)C(CC1=CCC2CC1C2(C)C)(NS(C)(=O)=O)C(F)(F)F

Structure B: COC(=O)C(CC1=CCC2CC1C2(C)C)(NS(C)(=O)=O)C(F)(F)F |t:6|

Perfect search for A or B returns A.

Exact search for A or B returns B.

 

Structure C: CC=C(C)C(=O)OCC(=CCO)C(=O)OC1CC(=C)C2CC=C(C)C2C2OC(=O)C(=C)C12

Exact search for C returns nothing. Perfect search for C returns C.

 

Is this a bug? 

Hi Imants,

 

I could not reproduce this.

I tried pasting it into Marvin in our JSP example: http://www.chemaxon.com/jchem/examples/db_search/index.jsp

(JChem 5.3.8) - The structures are still there in table "editexample".

I also tried directly using smiles, from jcsearch command line. (JChem 5.3.2)

 

A and B both retrieved A and B using full and duplicate searching, and also C always retrieved C.

 

Which version did you use? Do you have a custom standardization on your table?

 

Thanks,

Szabolcs

I just paste smiles above into the editor. And this is what I get.

Structure A exact match query (finds Structure B)

Structure A perfect match query (finds Structure A)

Structure B exact match query (finds Structure B)

Structure B perfect match query (finds Structure A)

 

We do fairly basic standardization. I doubt it is standardization related. 

Structure B was added to the database years ago with a version that is now very old. The Structure A was added with the version 5.3.3 (our current version).

 

Szabolcs,

I have a list of structures where I have the same problem as with Structure C. They all have the vague cis/trans double bond attachments. Perhaps this is how it should be? If the structure in the database has vague bonds then exact search with vague bonds does not find this structure, only perfect match does?

Imants

Hi,

I checked JChem 5.3.3 and the comming 5.4 version.

5.3.3 and 5.3.8: Structure C works correctly for me: finds itself for substructure, full and duplicate searches

Structure A and B doesn't find anything in substructure and full searches, but finds itself in duplicate search. In memory all searches run fine.

This is a bug it should find itself in the other two types as well.

However I have a different error than you. Can you confirm it's really C and not A or B that works incorrectly?

5.4: All searches work correctly.

I will check which change corrected the behaiour and what was the source of the problem.

Bye,

Robert

Hi,

 

We suggest upgrading jchem  to some version >=5.4, where these searches run correctly.

Robert