[IMPALA-819] Incorrect result in aggregate query with "where" based on inline view

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: Impala 1.3
• Fix Version/s: Impala 1.3
• Assignee: Unassigned
• Reporter: casey
• Created: 19/Feb/14 06:35
• Updated: 02/Mar/14 05:24
• Resolved: 02/Mar/14 05:24

Description

This happens on 1.2.4, 1.3.0, and master.

The result of the first query below should be 1. Somehow including the "where" clause causes a problem.

[localhost:21000] > select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1 WHERE c > 1;
Query: select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1 WHERE c > 1
+----------+
| count(*) |
+----------+
| 0        |
+----------+
Returned 1 row(s) in 0.12s

[localhost:21000] > explain select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1 WHERE c > 1;
Query: explain select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1 WHERE c > 1
+----------------------------------------------------------+
| Explain String                                           |
+----------------------------------------------------------+
| Estimated Per-Host Requirements: Memory=74.00MB VCores=1 |
|                                                          |
| 02:AGGREGATE [FINALIZE]                                  |
| |  output: count(*)                                      |
| |                                                        |
| 04:AGGREGATE [MERGE FINALIZE]                            |
| |  output: sum(count(*))                                 |
| |  having: count(*) > 1                                  |
| |                                                        |
| 03:EXCHANGE [PARTITION=UNPARTITIONED]                    |
| |                                                        |
| 01:AGGREGATE                                             |
| |  output: count(*)                                      |
| |                                                        |
| 00:SCAN HDFS [functional.alltypesagg]                    |
|    partitions=10/10 size=743.67KB                        |
+----------------------------------------------------------+
Returned 16 row(s) in 0.02s


[localhost:21000] > select * FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1 WHERE c > 1;
Query: select * FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1 WHERE c > 1
+-------+
| c     |
+-------+
| 10000 |
+-------+
Returned 1 row(s) in 0.13s


[localhost:21000] > select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1;
Query: select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1
+----------+
| count(*) |
+----------+
| 1        |
+----------+
Returned 1 row(s) in 0.12s

[localhost:21000] > explain select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1;
Query: explain select count(*) FROM ( SELECT count(*) AS c FROM alltypesagg) AS t1
+----------------------------------------------------------+
| Explain String                                           |
+----------------------------------------------------------+
| Estimated Per-Host Requirements: Memory=74.00MB VCores=1 |
|                                                          |
| 02:AGGREGATE [FINALIZE]                                  |
| |  output: count(*)                                      |
| |                                                        |
| 04:AGGREGATE [MERGE FINALIZE]                            |
| |  output: sum(count(*))                                 |
| |                                                        |
| 03:EXCHANGE [PARTITION=UNPARTITIONED]                    |
| |                                                        |
| 01:AGGREGATE                                             |
| |  output: count(*)                                      |
| |                                                        |
| 00:SCAN HDFS [functional.alltypesagg]                    |
|    partitions=10/10 size=743.67KB                        |
+----------------------------------------------------------+
Returned 15 row(s) in 0.02s