AnsweredAssumed Answered

MapR hive vs Apache hive

Question asked by gabi_k on Aug 9, 2013
Latest reply on Aug 17, 2013 by Ted Dunning
hi, 
i am referring to [THIS][1] 

To fix it i got APache HIVE version 0.7.1 (i am running MapR with Hive 0.7.1), and compiled it with the fix of the join problem (changed int into long variables).

after making some QA i find out that the Versions not match to MapR version!

for example: 
to load data with MapR hive i am using : _"LOAD DATA INPATH '/tmp/qa/new\_hive/123456.csv' OVERWRITE INTO TABLE table;"_ successfully. 
with the new hive-exec-0.7.1.jar i built, when i am running the exact same LOAD, i am getting: "FAILED: Error in semantic analysis: Line 1:17 Invalid path '/tmp/qa/new_hive/123456.csv': only "file" or "hdfs" file systems accepted"

it means that you have differences in the HIVE versions.. am i right?

How can i get the MapR HIVE code, so i could fix the JOIN bug?

Thanks,
Gabi.

  [1]: http://answers.mapr.com/questions/6807/wrong-results-when-a-partition-has-to-many-rows

Outcomes