Presently, QTL objects are resolved by the LookupExtractor one key at a time. This proposal is to support a vectorized query model for lookup keys. LookupExtractor presently has an applyAll() method which would be utilized for this case. Based on the LookupExtractor implementation, this would increase the query execution performance for lookups. For the existing design to support this functionality, it would require a configurable number of rows to be buffered before the lookups are resolved.
Presently, QTL objects are resolved by the LookupExtractor one key at a time. This proposal is to support a vectorized query model for lookup keys. LookupExtractor presently has an applyAll() method which would be utilized for this case. Based on the LookupExtractor implementation, this would increase the query execution performance for lookups. For the existing design to support this functionality, it would require a configurable number of rows to be buffered before the lookups are resolved.