[DRILL-1062] DRILL does not handle NULLS FIRST/LAST correctly in ORDER BY clause

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.8.0
• Component/s: Functions - Drill
• Assignee: Chun Chang
• Reporter: Jinfeng Ni
• Created: 24/Jun/14 21:08
• Updated: 29/Apr/15 22:14
• Resolved: 24/Feb/15 06:37

Description

ORDER BY clause could specify nulls first or nulls last. Currently, DRILL will always use nulls last policy.

select tbl.topping[3].type from dfs.`/Users/jni/work/incubator-drill/exec/ref/target/test-classes/donuts.json` as tbl order by 1 nulls last;
------------

EXPR$0

------------

Chocolate

Maple

Powdered Sugar

Powdered Sugar

null

------------
5 rows selected (0.156 seconds)
0: jdbc:drill:zk=local> select tbl.topping[3].type from dfs.`/Users/jni/work/incubator-drill/exec/ref/target/test-classes/donuts.json` as tbl order by 1 nulls first;
------------

EXPR$0

------------

Chocolate

Maple

Powdered Sugar

Powdered Sugar

null

------------
5 rows selected (0.186 seconds)

Issue Links

• incorporates: DRILL-1763 Add exclusion of Emacs TAGS files in .gitignore and RAT configuration
• is duplicated by: DRILL-1156 mondrian117.q - result set with group by stitched wrong