[ https://issues.apache.org/jira/browse/HIVE-20437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16606253#comment-16606253 ]
Janaki Lahorani commented on HIVE-20437: ---------------------------------------- The following test failure is not related to this patch: org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver[allcolref_in_udf] (batchId=56) It can be seen for other runs as well. Also reproduces without this patch. > Handle schema evolution from float, double and decimal > ------------------------------------------------------ > > Key: HIVE-20437 > URL: https://issues.apache.org/jira/browse/HIVE-20437 > Project: Hive > Issue Type: Bug > Reporter: Janaki Lahorani > Assignee: Janaki Lahorani > Priority: Major > Attachments: HIVE-20437.1.patch, HIVE-20437.2.patch, > HIVE-20437.3.patch > > > When data created as float, double or decimal in parquet format is read back > using some other type, errors are seen. Parquet should behave just like any > other format. If the value is valid for the new type, data is retuned > otherwise null has to be returned. -- This message was sent by Atlassian JIRA (v7.6.3#76005)