[HIVE-3577] Subquery Unnesting

• Type: Sub-task
• Status: Open
• Priority: Critical
• Resolution: Unresolved
• Affects Version/s: 0.10.0, 0.11.0, 0.12.0, 0.13.0
• Labels: SQL
• Assignee: Unassigned
• Reporter: Shane Huang
• Created: 15/Oct/12 06:51
• Updated: 25/Dec/13 22:43

Description

Hive requires subquery to be used only in the FROM clause. But real-world queries can be much more complex and subqueries could appear in WHERE, having and SELECT clauses. Many of such subqueries can be transformed/unnested to various join operations. This is an umbrella id to include all subquery unnesting work, specifically:

1. subquery in WHERE conditions, including subquery used as operand of IN, EXISTS, ISNULL ALL/ANY/SOME, and etc.

2. correlated subquery (a correlated subquery is a subquery that refers to a column of a table not in its FROM clause),e.g.

select * from t1 where exists ( select * from t2 where t1.b = t2.y );

3. scalar subquery (a subquery that returns exactly one column value from one row). e.g.

select a,b,c,d,e,(select z from t2 where t2.y = t1.b and z != 99 ) from t1;

4. top-level subquery, e.g.

(select * from t1) union all (select * from t2) union (all select * from t3 order by 1);

Issue Links

• is related to: HIVE-784 Support uncorrelated subqueries in the WHERE clause