[IMPALA-1629] Column stats for CHAR and VARCHAR not populated by COMPUTE STATS.

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 2.0.1, Impala 2.1
• Fix Version/s: Impala 2.2
• Labels: compute-stats
• Assignee: Alexander Behm
• Reporter: Alexander Behm
• Created: 22/Dec/14 21:39
• Updated: 20/Feb/15 00:55
• Resolved: 20/Feb/15 00:55

Description

Even if COMPUTE STATS is run, the column stats for CHAR/VARCHAR column remain unpopulated. Follow these steps for a repro:

create table test (x int, y varchar(20), z char(10));
insert into test values(1, cast('aaa' as varchar(20)), cast('bbb' as char(10)));
compute stats test;
[localhost:21000] > show column stats test;
Query: show column stats test
+--------+-------------+------------------+--------+----------+----------+
| Column | Type        | #Distinct Values | #Nulls | Max Size | Avg Size |
+--------+-------------+------------------+--------+----------+----------+
| x      | INT         | 1                | -1     | 4        | 4        |
| y      | VARCHAR(20) | -1               | -1     | -1       | -1       |
| z      | CHAR(10)    | -1               | -1     | 10       | 10       |
+--------+-------------+------------------+--------+----------+----------+