[HIVE-15251] Provide support for complex expressions in ON clauses for OUTER joins

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.2.0
• Fix Version/s: 2.2.0
• Component/s: CBO, Parser
• Assignee: Jesus Camacho Rodriguez
• Reporter: Jesus Camacho Rodriguez
• Created: 21/Nov/16 18:57
• Updated: 26/Jul/17 00:03
• Resolved: 06/Dec/16 00:28

Description

Follow-up of HIVE-15211.

Currently, we have some restrictions on the predicates that we can use in ON clauses for outer joins.

This patch is an extension to overcome these restrictions. Follow-up work will focus on identifying some cases, such as disjunctive predicates, that can be run more efficiently than with a cartesian product.

It will allow to write queries that currently fail in Hive such as:

-- Disjunctions
SELECT *
FROM src1 LEFT OUTER JOIN src
ON (src1.key=src.key
  OR src1.value between 100 and 102
  OR src.value between 100 and 102)
LIMIT 10;
-- Conjunction with multiple inputs references in one side
SELECT *
FROM src1 RIGHT OUTER JOIN src
ON (src1.key+src.key >= 100
  AND src1.key+src.key <= 102)
LIMIT 10;
-- Conjunct with no references
SELECT *
FROM src1 FULL OUTER JOIN src
ON (src1.value between 100 and 102
  AND src.value between 100 and 102
  AND true)
LIMIT 10;

Issue Links

• breaks: HIVE-15369 Extend column pruner to account for residual filter expression in Join operator
• breaks: HIVE-15370 Include Join residual filter expressions in user level EXPLAIN
• is blocked by: HIVE-15327 Outerjoin might produce wrong result depending on joinEmitInterval value
• relates to: HIVE-15211 Provide support for complex expressions in ON clauses for INNER joins
• links to: RB link