It looks like a binary issue, which is really bizarre given that all the code is in the same jar, and unpacking shows nothing suspicious. There’s another switch that handles this token before the main switch. That switch has a case for TOK_FILEFORMAT_GENERIC, but that never gets executed (if it did, the code wouldn’t go into the 2nd switch), and then the 2nd switch goes to the default path and complains it cannot handle TOK_FILEFORMAT_GENERIC (which it cannot, indeed). Rebuilding and replacing jar made the error go away on my machine. I’ll redo the RC...
On 16/2/8, 13:06, "Sergey Shelukhin" <ser...@hortonworks.com> wrote: >I can repro. Not sure why the main issue is happening. >The AST for create table in master is >TOK_CREATETABLE > TOK_TABNAME > moooo > TOK_LIKETABLE > TOK_TABCOLLIST > TOK_TABCOL > i > TOK_INT > TOK_FILEFORMAT_GENERIC > orc > >Which should produce the same error in this switch statement, but it >doesn’t, which makes me thing it doesn’t normally execute for this query. >Moreover, q file tests as of the RC commit on branch-2.0 can create a >table with this query just fine. > >From: Prasanth Jayachandran ><pjayachand...@hortonworks.com<mailto:pjayachand...@hortonworks.com>> >Reply-To: "dev@hive.apache.org<mailto:dev@hive.apache.org>" ><dev@hive.apache.org<mailto:dev@hive.apache.org>> >Date: Monday, February 8, 2016 at 12:53 >To: "dev@hive.apache.org<mailto:dev@hive.apache.org>" ><dev@hive.apache.org<mailto:dev@hive.apache.org>> >Subject: Re: [VOTE] Apache Hive 2.0.0 Release Candidate 1 > >https://issues.apache.org/jira/browse/HIVE-13023
