All tables needed updating in new update

Hi

I installed the updates to IJC 5.6 yesterday and again needed to update all databases.  This was slightly faster than last time but still had all four cores working for over six hours.  This seems a regular feature of the upgrades recently - is this likely to continue ?

I have another question about the upgrade / updates to the databases.  Are any of the columns actually recalculated, so if I go back, will I get different result ?  If this is the case will there be any memory of which columns were calculated previously or is it a case of the programme looking for the correct column headers ?  I have some tables in which I've cLogP values from four different methods and wouldn't like to think that the wrong column(s) have been recalculated.

Any thoughts ?

Thanks

Steve

 

I installed the updates to IJC 5.6 yesterday and 

again needed to update all databases.  This was slightly faster than 

last time but still had all four cores working for over six hours.  This

 seems a regular feature of the upgrades recently - is this likely to 

continue ?

Regenerting the tables will often be needed for major releases. Occasionally it is needed even for minor releases.

The reasons are various, but generally in two categories:

1. when details of the search agorithm are improved, osmetimes needing a change to the descriptors (fingerprints). In this case the descripters need to be refreshed so that they correspond to what the seach algorithm needs.

2. improvents in the property predictors e.g. a better logP calcualtor. In these cases the values need to be updated so that they are the improved values.

We try to minimise the need for re-generation as far as possible, and whether re-generation is needed is mentioned in the release notes:
http://www.chemaxon.com/instantjchem/Release_Notes_Cumulative.html

Are any of the columns actually recalculated, so 

if I go back, will I get different result ?  If this is the case will 

there be any memory of which columns were calculated previously or is it

 a case of the programme looking for the correct column headers ?  I 

have some tables in which I've cLogP values from four different methods 

and wouldn't like to think that the wrong column(s) have been 

recalculated.

Yes, calculated columns (chemical terms fields) will be recalculated, but the details of the calculation should remain the same. Just that the values may be slightly more correct after re-calculation if the algorithm has been improved.

Tim