[JENA-1440] Map ByteBuffer direct to NodeIds, avoiding a Record object (TupleIndexRecord)

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Done
• Affects Version/s: Jena 3.5.0, Jena 3.6.0
• Fix Version/s: Jena 3.6.0
• Component/s: TDB2
• Assignee: Andy Seaborne
• Reporter: Andy Seaborne
• Created: 27/Nov/17 18:32
• Updated: 17/Dec/17 19:46
• Resolved: 04/Dec/17 14:42

Description

Avoiding going bytes-> Record->Tuple<NodeId> can save about 25% of time in a large index scan (but there are lots of other things going on in a typical query and this change would be negligible).

Issue Links

• links to: GitHub Pull Request #317