Inconsistent Results

User 2a533dbb3b

15-05-2009 20:19:33

I seem to be experiencing inconsistent search results.  The below is four successive searches using the same instance of JChemSearch.  The 2nd, 3rd and 4th structures are the same as you can tell by the Query value.  The problem is that the 2nd received no hits and the 3rd and 4th both received hits.  Obviously, we wouldn't search the same structure 3 times but I am experiencing the behavior where somtime the structure brings back a result and other times it does not.  Any idea what might be causing this?


Thanks





Fri May 15 16:02:53 EDT 2009
Search mode: EXACT
Structure table: CATALOG.CHEMSTRUCTURE
Query: [H]OC(=O)C1=C(C(=O)O[H])C(Cl)=C(Cl)C(Cl)=C1Cl.[H]OC(=O)C2=C(C(=O)O[H])C(Cl)=C(Cl)C(Cl)=C2Cl.[H]OC3=C(C([H])=C(C([H])=C3C([H])([H])N(C([H])([H])[H])C([H])([H])[H])C([H])([H])N(C([H])([H])[H])C([H])([H])[H])C([H])([H])N(C([H])([H])[H])C([H])([H])[H]
Screened: 0
Hits: 0
Cache loading: 30124 ms
Cache size (this table / total): 3.57 / 3.57 MBytes
Total time: 1017 ms  Screening: 16 ms
Processing threads: 4
Current / peak / maximum searches per minute: 1 / 1 / Unlimited


Fri May 15 16:03:23 EDT 2009
Search mode: EXACT
Structure table: CATALOG.CHEMSTRUCTURE
Query: [H]OC(=O)C1=C([H])C([H])=C(C([H])=C1[H])C(=O)O[H]
Screened: 2
Hits: 0
Total time: 109 ms  Screening: 0 ms
Processing threads: 4
Current / peak / maximum searches per minute: 2 / 2 / Unlimited


Fri May 15 16:03:23 EDT 2009
Search mode: EXACT
Structure table: CATALOG.CHEMSTRUCTURE
Query: [H]OC(=O)C1=C([H])C([H])=C(C([H])=C1[H])C(=O)O[H]
Screened: 2
Hits: 1
Total time: 109 ms  Screening: 15 ms
Processing threads: 4
Current / peak / maximum searches per minute: 3 / 3 / Unlimited


Fri May 15 16:03:24 EDT 2009
Search mode: EXACT
Structure table: CATALOG.CHEMSTRUCTURE
Query: [H]OC(=O)C1=C([H])C([H])=C(C([H])=C1[H])C(=O)O[H]
Screened: 2
Hits: 1
Total time: 110 ms  Screening: 0 ms
Processing threads: 4
Current / peak / maximum searches per minute: 4 / 4 / Unlimited

ChemAxon 9c0afc9aaf

15-05-2009 20:46:46

The 2nd, 3rd and 4th structures are the same as you can tell by the Query value.


These SMARTS representations are used here because they are compact and fit well here.


However they cannot store all structural information, for example enchanced stereo representation in MDL SDfiles.


So they are there just to provide some clue about the query structure, but cannot be used to decide if a search returned the correct results or not.


I suggest to check your original query structures, I suspect they are not exactly the same, and the search works as required.


If you think you have found a search error please send us the query and target which does not (always) find each other.


Best regards,


Szilard


 

ChemAxon a3d59b832c

01-06-2009 09:45:32

Just for the record:


This conversation was continued in emails. Finally it turned out, that the setTimeout() property was used and depending on the load on the machine, the searching may or may not
reached certain hits before the timeout.


This page describes how the timeout status can be queried from JChem API:


http://www.chemaxon.com/jchem/doc/guide/search/index.html#sss_retrieve