[KYLIN-1453] cuboid sharding based on specific column

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: v1.5.2
• Assignee: Hongbin Ma
• Reporter: Hongbin Ma
• Created: 26/Feb/16 07:44
• Updated: 26/May/16 09:16
• Resolved: 21/Apr/16 11:21

Description

currently cuboid sharding is based on the hash value of the row key. If we allow computing hash value on a specific column C, then we're actually partitioning the data w.r.t C. The benefit is that if we later a query with filter like "where C = 'xyz'", kylin can skip shards. Also for filter like "where C IN

{many candidates}

", kylin can prune candidates sent to different shards

Issue Links

• is related to: KYLIN-1428 Scalable dictionary to support cardinality of billions