hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Manuel Meßner (JIRA) <j...@apache.org>
Subject [jira] [Created] (HIVE-9595) Test whether schema evolution using parquet is possible
Date Fri, 06 Feb 2015 08:14:35 GMT
Manuel Meßner created HIVE-9595:
-----------------------------------

             Summary: Test whether schema evolution using parquet is possible
                 Key: HIVE-9595
                 URL: https://issues.apache.org/jira/browse/HIVE-9595
             Project: Hive
          Issue Type: Test
          Components: File Formats, Tests
            Reporter: Manuel Meßner
            Priority: Trivial


With https://issues.apache.org/jira/browse/HIVE-7554 a unfortunate else branch was introduced
to org.apache.hadoop.hive.ql.io.parquet.read.DataWritableReadSupport . That else branch fired
an exception in case the parquet file schema had less fields than the table schema. So adding
columns to a parquet backed table and query the added columns on old partitions wasn't possible
anymore. Which was very unfortunate for my team. https://issues.apache.org/jira/browse/HIVE-7800
fixed that, see point 3 in Daniel Weeks' comment https://issues.apache.org/jira/browse/HIVE-7800?focusedCommentId=14158594&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14158594
.

However Hive-7800 did not introduce a test case for that. I'd love to see one.

Best,
Manuel




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

Mime
View raw message