drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jason Altekruse (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DRILL-1664) Drill gives wrong count on a parquet file which is created as a table by drill
Date Thu, 13 Nov 2014 23:27:34 GMT

    [ https://issues.apache.org/jira/browse/DRILL-1664?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14211467#comment-14211467
] 

Jason Altekruse commented on DRILL-1664:
----------------------------------------

This is a bug in our optimized path for using the parquet meta-data to return results for
a count operation without a full scan. This is currently failing on repeated data (all data
read out of csv files is read as a repeated varchar). A fix will be posted shortly.

> Drill gives wrong count on a parquet file which is created as a table by drill
> ------------------------------------------------------------------------------
>
>                 Key: DRILL-1664
>                 URL: https://issues.apache.org/jira/browse/DRILL-1664
>             Project: Apache Drill
>          Issue Type: Bug
>            Reporter: Venkata krishnan Sowrirajan
>
> Steps carried out:
> 0: jdbc:drill:> select * from `t2.csv`;
> +------------+
> |  columns   |
> +------------+
> | ["9711942","HX362083"] |
> | ["9707867","HX357851"] |
> +------------+
> 2 rows selected (0.123 seconds)
> 0: jdbc:drill:> create table `t2-csv` as select * from `t2.csv`;
> +------------+---------------------------+
> |  Fragment  | Number of records written |
> +------------+---------------------------+
> | 0_0        | 2                         |
> +------------+---------------------------+
> 1 row selected (0.252 seconds)
> 0: jdbc:drill:> select * from `t2-csv`;
> +------------+
> |  columns   |
> +------------+
> | ["9711942","HX362083"] |
> | ["9707867","HX357851"] |
> +------------+
> 2 rows selected (0.116 seconds)
> 0: jdbc:drill:> select count(*) from `t2-csv`
> . . . . . . . > ;
> +------------+
> |   EXPR$0   |
> +------------+
> | 4          |
> +------------+
> 1 row selected (0.128 seconds)
> Is there a similar bug for this already filed? If there is a similar bug for this, mark
this as duplicate as I couldn't find that.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message