You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@drill.apache.org by "徐波 (JIRA)" <ji...@apache.org> on 2015/08/13 11:14:45 UTC

[jira] [Updated] (DRILL-3501) No vector has initialized when the value of qualifier is null while using Hbase storage

     [ https://issues.apache.org/jira/browse/DRILL-3501?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

徐波 updated DRILL-3501:
----------------------
    Component/s: Storage - HBase

> No vector has initialized when the value of qualifier is null while using Hbase storage
> ---------------------------------------------------------------------------------------
>
>                 Key: DRILL-3501
>                 URL: https://issues.apache.org/jira/browse/DRILL-3501
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Storage - HBase
>    Affects Versions: 0.9.0, 1.0.0, 1.1.0
>            Reporter: 徐波
>         Attachments: 0001-init-vector-in-setup.patch
>
>
> drill throws exception while querying from hbase:
> Failure while reading vector.  Expected vector class of org.apache.drill.exec.vector.NullableVarBinaryVector but was holding vector class org.apache.drill.exec.vector.NullableIntVector.
> My sql:
> select t2.A.A1, t2.A.B2 from hdfs.root.`table1` as t1 inner join hbase.`table2` as t2 on table1.c1=t2.c1
> Now the table2 has 3 regions and A.B2 on the last region is null while containing values on the other regions. The node creates a NullableIntVector while quering from the last region,but the nodes create a NullableVarBinaryVector while quering from the other regions.So drill throws Exception on the join node which get two type of vector for one column.
> I recommend that we should initialize vector for selected columns in the  method named 'org.apache.drill.exec.store.hbase.HBaseRecordReader.setup'



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