I am using HDP 2.5.
Getting following error while running select * from query with different user. like test user.
Error:- See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. Aug 28, 2017 7:40:44 AM WARNING: org.apache.parquet.CorruptStatistics: Ignoring statistics because created_by could not be parsed (see PARQUET-251): parquet-mr (build 32c46643845ea8a705c35d4ec8fc654cc8ff816d) org.apache.parquet.VersionParser$VersionParseException: Could not parse created_by: parquet-mr (build 32c46643845ea8a705c35d4ec8fc654cc8ff816d) using format: (.+) version ((.*) )?\(build ?(.*)\)
The underlying reason for this seems to that this file was written using different version of parquet-hadoop and trying to read with another version.
Look out for the hive version run the below command
# locate parquet-hadoop
Please upgrade the parquet version being used and retry
Please check if you have the "parquet-logging.properties" file present inside the "/etc/hive/conf/" directory or not?
If not then please copy that file "parquet-logging.properties" from hive2 conf directory to "/etc/hive/conf/" directory and then try again. That file might be missing there due to : https://issues.apache.org/jira/browse/AMBARI-21265
Is that file also available inside the "/etc/hive/conf" as well as inside "/etc/hive/conf/conf.server" ?
Also as we see that it is a WARNING message so we can suppress it as well.
Or else can you please share the complete stackTrace so that we can see the complete trace to find out the APIs involved in this WARN.
The writing and reading should use the same version of parquet-hadoop.Use the newer version of parquet-hadoop for writing parquet files Also check the permission on the parquet-0.log.lock