[
https://issues.apache.org/jira/browse/PHOENIX-1362?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14247701#comment-14247701
]
ASF GitHub Bot commented on PHOENIX-1362:
-----------------------------------------
GitHub user dhacker1341 opened a pull request:
https://github.com/apache/phoenix/pull/30
PHOENIX-1362 Fix Min/MaxAggregateFunction Aggregator max length
Turned out the expression compiler was seeing the max length of 0 on fixed
length data types because getMaxLength() on the aggregator was not being
overridden.
Let me know if I need to change anything especially on the test side. Not
too familiar yet with the style of the project.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/dhacker1341/phoenix master
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/phoenix/pull/30.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #30
----
commit 489f57a93a734b7d47aed864ea8f71d6bc01d0d7
Author: David <[email protected]>
Date: 2014-12-16T03:46:17Z
PHOENIX-1362 Fix Min/MaxAggregateFunction Aggregator max length
----
> Min/max query on CHAR columns containing values with '0' as prefix always
> returns null
> --------------------------------------------------------------------------------------
>
> Key: PHOENIX-1362
> URL: https://issues.apache.org/jira/browse/PHOENIX-1362
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.1
> Environment: HBase 0.98.4
> RHEL 6.5
> Reporter: Hari Krishna Dara
> Assignee: Dave Hacker
> Labels: aggregate, char
>
> - Create a table with CHAR type and insert a few strings that start with 0.
> - Select min()/max() on the column, you always get null value.
> {noformat}
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> create table TT(VAL1 integer not
> null, VAL2 char(2), val3 varchar, VAL4 varchar constraint PK primary key
> (VAL1));
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> upsert into TT values (0, '00', '00',
> '0');
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> upsert into TT values (1, '01', '01',
> '1');
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> upsert into TT values (2, '02', '02',
> '2');
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> select * from TT;
> +------------+------+------------+------------+
> | VAL1 | VAL2 | VAL3 | VAL4 |
> +------------+------+------------+------------+
> | 0 | 00 | 00 | 0 |
> | 1 | 01 | 01 | 1 |
> | 2 | 02 | 02 | 2 |
> +------------+------+------------+------------+
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> select min(VAL1), max(VAL1) from TT;
> +------------+------------+
> | MIN(VAL1) | MAX(VAL1) |
> +------------+------------+
> | 0 | 2 |
> +------------+------------+
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> select min(VAL2), max(VAL2) from TT;
> +------------+------------+
> | MIN(VAL2) | MAX(VAL2) |
> +------------+------------+
> | null | null |
> +------------+------------+
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> select min(VAL3), max(VAL3) from TT;
> +------------+------------+
> | MIN(VAL3) | MAX(VAL3) |
> +------------+------------+
> | 00 | 02 |
> +------------+------------+
> 0: jdbc:phoenix:isthbase01-mnds2-1-crd> select min(VAL4), max(VAL4) from TT;
> +------------+------------+
> | MIN(VAL4) | MAX(VAL4) |
> +------------+------------+
> | 0 | 2 |
> +------------+------------+
> {noformat}
> As you can see, the query on VAL2 which is of type CHAR(2) returns null,
> while the same exact values on VAL3 which is of type VARCHAR work as expected.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
