Chun Chang created DRILL-2581:
---------------------------------
Summary: count(*) on a single wide record file took too long
Key: DRILL-2581
URL: https://issues.apache.org/jira/browse/DRILL-2581
Project: Apache Drill
Issue Type: Bug
Components: Execution - Data Types
Affects Versions: 0.8.0
Reporter: Chun Chang
Assignee: Hanifi Gunes
This is with the 0.8 release candidate.
I have a file containing a single record. The file is about 3.5 MB in size. A
count(*) on the file took over one and half minute.
{code}
0: jdbc:drill:schema=dfs.drillTestDirComplexJ> select * from sys.version;
+------------+----------------+-------------+-------------+------------+
| commit_id | commit_message | commit_time | build_email | build_time |
+------------+----------------+-------------+-------------+------------+
| 462e50ce9c4b829c2a4bafdeb9763bfba677c726 | DRILL-2575:
FragmentExecutor.cancel() blasts through state transitions regardless of
current state | 25.03.2015 @ 21:11:23 PDT | [email protected] | 26.03.2015 @
10:51:15 PDT |
+------------+----------------+-------------+-------------+------------+
1 row selected (0.07 seconds)
0: jdbc:drill:schema=dfs.drillTestDirComplexJ> select count(*) from
`singlewide.json`;
+------------+
| EXPR$0 |
+------------+
| 1 |
+------------+
1 row selected (99.91 seconds)
{code}
The data file can be downloaded shortly from
http://apache-drill.s3.amazonaws.com/list.html
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
