[ https://issues.apache.org/jira/browse/IMPALA-1871?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jim Apple updated IMPALA-1871: ------------------------------ Attachment: (was: tempfile.txt) > Impala is unable to read timestamp field value in avro written by hive > ---------------------------------------------------------------------- > > Key: IMPALA-1871 > URL: https://issues.apache.org/jira/browse/IMPALA-1871 > Project: IMPALA > Issue Type: Bug > Components: Backend > Affects Versions: Impala 2.1.1, Impala 2.3.0 > Environment: CDH 5.4 > Reporter: Syed A. Hashmi > Assignee: Dan Hecht > Attachments: tempfile.txt > > > Hive recently added support for writing timestamp values in avro format. > Field description looks like this: > {code} > { "type": "long", "logicalType": "timestamp-millis" } > {code} > I have created an avro table with data using hive. When I am trying to read > this field from Impala, it reads as garbage, while from hive I can read it > fine. Here is the result from hive: > {code} > 0: jdbc:hive2://shashmi-test-1.vpc.cloudera.c> select timestamp_field from > hiveclient_hiveclient_uncompressed; > +-------------------------+--+ > | timestamp_field | > +-------------------------+--+ > | 2012-02-21 07:08:09.12 | > | 1998-02-21 07:08:09.12 | > | 1970-01-01 00:00:01.0 | > | 2038-01-19 03:14:07.0 | > | NULL | > +-------------------------+--+ > {code} > While the output from Impala is incorrect: > {code} > [shashmi-test-2.vpc.cloudera.com:21000] > select timestamp_field from > hiveclient_hiveclient_uncompressed; > Query: select timestamp_field from hiveclient_hiveclient_uncompressed > +-----------------+ > | timestamp_field | > +-----------------+ > | 1329836889120 | > | 888073689120 | > | 28801000 | > | 2147512447000 | > | NULL | > +-----------------+ > {code} -- This message was sent by Atlassian JIRA (v6.3.15#6346)