Error creating Bingo molecule index

[Mitch Miller]

Hello,

I'm trying to create an index on a 400k record table of molfiles using this command:

   create index chem_mol_idx on chem using bingo_idx (molecule bingo.molecule)

and the following error appears:

 NOTICE: bingo.index: 72000 structures processed ERROR: error: bingo: Error while executing build index procedure error: bingo buffer cache binary data: internal error: can not add cmf to the cache because is not enough space SQL state: XX000  

And the error indicates that one (or more) of the structures in the table is too large.

How can I find the record(s) with the molfiles causing the error?

Mitch

[Savelyev Alexander]

Hi Mitch

Could you please try to set log level to "DEBUG1"? It should output structure number in section (each section contains 64k structures) so it will be easier to find the huge structure. 

With best regards

Aleksandr

[Mitch Miller]

Thanks, Aleksandr,

How do I set the log level?

Mitch

[Savelyev Alexander]

One can change log levels using PostgreSQL settings

https://www.postgresql.org/docs/current/runtime-config-logging.html

Changing  logging_collector , log_min_messages , log_min_error_statement should be enough

Another possible way to find a huge structure is to find length of "CompactMolecule"

https://lifescience.opensource.epam.com/bingo/user-manual-postgres.html#conversion-to-binary-format

The function should return cmf (internal format), therefore assuming max block size is 8kb then the following SQL can help

select *  FROM $table;  

WHERE octet_length(bingo.CompactMolecule($column, 0) ) > 8000

Thanks

Aleskandr

 

[Mitch Miller]

Thanks,  Aleksandr!!

 

>>select *  FROM $table;  

>>WHERE octet_length(bingo.CompactMolecule($column, 0) ) > 8000

allowed me to remove the large molfiles from my table and create the index.

Cheers,

Mitch