[IMPALA-2446] Incorrect result (left join)

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 2.3.0
• Fix Version/s: Impala 2.3.0, Impala 2.2.9
• Labels: correctness nested_types query_generator
• Target Version: Impala 2.3.0, Impala 2.2.9
• Assignee: Dimitris Tsirogiannis
• Reporter: Taras Bobrovytsky
• Created: 29/Sep/15 19:59
• Updated: 15/Dec/15 22:00
• Resolved: 06/Oct/15 16:00

Description

Query:

SELECT t3976.field_13, t3977.pos
FROM table_0 t3976
LEFT JOIN t3976.field_0.field_5 t3977
WHERE (t3976.field_13) = (t3977.pos)

Plan:

+------------------------------------------------------------------------------------+
| Explain String                                                                     |
+------------------------------------------------------------------------------------+
| Estimated Per-Host Requirements: Memory=96.00MB VCores=1                           |
| WARNING: The following tables are missing relevant table and/or column statistics. |
| random_nested_db_0.table_0                                                         |
|                                                                                    |
| 05:EXCHANGE [UNPARTITIONED]                                                        |
| |                                                                                  |
| 01:SUBPLAN                                                                         |
| |                                                                                  |
| |--04:NESTED LOOP JOIN [RIGHT OUTER JOIN]                                          |
| |  |  join predicates: (t3977.pos) = (t3976.field_13)                              |
| |  |                                                                               |
| |  |--02:SINGULAR ROW SRC                                                          |
| |  |                                                                               |
| |  03:UNNEST [t3976.field_0.field_5 t3977]                                         |
| |                                                                                  |
| 00:SCAN HDFS [random_nested_db_0.table_0 t3976]                                    |
|    partitions=1/1 files=1 size=32.98MB                                             |
+------------------------------------------------------------------------------------+

The first few rows look like this:

+-------------+------+
| field_13    | pos  |
+-------------+------+
| -756662132  | NULL |
| 441414792   | NULL |
| -1783035865 | NULL |
| -714649183  | NULL |
| -1337026208 | NULL |
| -1754382556 | NULL |
| -382183064  | NULL |
| -838360500  | NULL |
| -1731241459 | NULL |
| 692876994   | NULL |

To access the machine with the loaded data:

ssh [e-mail] -p 33334 (pw: cloudera)
run-all.sh && start-impala-cluster.py
impala-shell.sh
use random_nested_db_0;

(core dumps are saved to /tmp/core_files)