[IMPALA-2088] Incorrect result (Analytic Functions)

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 2.3.0
• Fix Version/s: Impala 2.2.7, Impala 2.3.0
• Labels: correctness query_generator
• Target Version: Impala 2.2.7, Impala 2.3.0
• Assignee: Alexander Behm
• Reporter: Taras Bobrovytsky
• Created: 22/Jun/15 19:35
• Updated: 01/Sep/15 23:53
• Resolved: 27/Jul/15 20:35

Description

The query should clearly return 0 rows, however 2 rows are returned.

Query:

SELECT
LEAD(-496, 81) OVER (ORDER BY t1.double_col DESC, t1.id ASC) AS int_col
FROM alltypestiny t1
WHERE
5 = 6
UNION
SELECT
794.67 AS decimal_col
FROM alltypes t1
WHERE
5 = 6

Result:

+---------+
| int_col |
+---------+
| NULL    |
| 794.67  |
+---------+

Plan:

+--------------------------------------------------------------+
| Explain String                                               |
+--------------------------------------------------------------+
| Estimated Per-Host Requirements: Memory=170.00MB VCores=2    |
|                                                              |
| 10:EXCHANGE [UNPARTITIONED]                                  |
| |                                                            |
| 09:AGGREGATE [FINALIZE]                                      |
| |  group by: int_col                                         |
| |                                                            |
| 08:EXCHANGE [HASH(int_col)]                                  |
| |                                                            |
| 05:AGGREGATE                                                 |
| |  group by: int_col                                         |
| |                                                            |
| 00:UNION                                                     |
| |                                                            |
| |--04:SCAN HDFS [functional.alltypes t1]                     |
| |     partitions=24/24 files=24 size=478.45KB                |
| |                                                            |
| 07:EXCHANGE [RANDOM]                                         |
| |                                                            |
| 03:ANALYTIC                                                  |
| |  functions: lead(-496, 81, NULL)                           |
| |  order by: double_col DESC, id ASC                         |
| |  window: ROWS BETWEEN UNBOUNDED PRECEDING AND 81 FOLLOWING |
| |                                                            |
| 06:MERGING-EXCHANGE [UNPARTITIONED]                          |
| |  order by: double_col DESC, id ASC                         |
| |                                                            |
| 02:SORT                                                      |
| |  order by: double_col DESC, id ASC                         |
| |                                                            |
| 01:SCAN HDFS [functional.alltypestiny t1]                    |
|    partitions=4/4 files=4 size=460B                          |
+--------------------------------------------------------------+