[jira] [Updated] (DRILL-5018) Metadata cache has duplicate columnTypeInfo values

["Robert Hou (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/DRILL-5018?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Robert Hou updated DRILL-5018:
------------------------------
    Attachment: lineitem_999.parquet
                lineitem_1_0_1.parquet

I have attached the parquet files for the first lineitem table described in the Description.  Create a year directory and then a month directory and put one of the files in the month directory.  Create a second year directory and month directory and put the second file in the month directory.

> Metadata cache has duplicate columnTypeInfo values
> --------------------------------------------------
>
>                 Key: DRILL-5018
>                 URL: https://issues.apache.org/jira/browse/DRILL-5018
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Metadata
>    Affects Versions: 1.8.0
>            Reporter: Robert Hou
>            Assignee: Parth Chandra
>         Attachments: lineitem_1_0_1.parquet, lineitem_999.parquet
>
>
> This lineitem table has duplicate entries in its metadata file, although the entries have slightly different values.  This lineitem table uses directory-based partitioning on year and month.
>   "columnTypeInfo" : {
>     "L_RETURNFLAG" : {
>       "name" : [ "L_RETURNFLAG" ],
>       "primitiveType" : "BINARY",
>       "originalType" : null,
>       "precision" : 0,
>       "scale" : 0,
>       "repetitionLevel" : 0,
>       "definitionLevel" : 1
>     },
>     "l_returnflag" : {
>       "name" : [ "l_returnflag" ],
>       "primitiveType" : "BINARY",
>       "originalType" : "UTF8",
>       "precision" : 0,
>       "scale" : 0,
>       "repetitionLevel" : 0,
>       "definitionLevel" : 0
>     },
> This lineitem table has two entries in its metadata file for each column, but the two entries have different column names (adding a zero).  It also has slightly different values.  This lineitem table was created using CTAS with the first table above.
>     "l_shipinstruct" : {
>       "name" : [ "l_shipinstruct" ],
>       "primitiveType" : "BINARY",
>       "originalType" : "UTF8",
>       "precision" : 0,
>       "scale" : 0,
>       "repetitionLevel" : 0,
>       "definitionLevel" : 0
>     },
>     "L_SHIPINSTRUCT0" : {
>       "name" : [ "L_SHIPINSTRUCT0" ],
>       "primitiveType" : "BINARY",
>       "originalType" : null,
>       "precision" : 0,
>       "scale" : 0,
>       "repetitionLevel" : 0,
>       "definitionLevel" : 1
>     },



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