Siddharth Seth created HIVE-12624:
-------------------------------------
Summary: HYBRID ORC split strategy can generate inconsistent
splits across two otherwise identical runs
Key: HIVE-12624
URL: https://issues.apache.org/jira/browse/HIVE-12624
Project: Hive
Issue Type: Bug
Reporter: Siddharth Seth
When using the HYBRID strategy to execute the same query twice - the raw splits
generated in the two runs can differ. Wondering if this has something to do
with footer caching.
cc [~prasanth_j]
14: [
org.apache.hadoop.hive.ql.io.orc.OrcInputFormat:hdfs:///...date=2015-11-03/000000_0:3+367753]
!= [
org.apache.hadoop.hive.ql.io.orc.OrcInputFormat:hdfs://...date=2015-11-03/000000_0:0+367958]
65: [
org.apache.hadoop.hive.ql.io.orc.OrcInputFormat:hdfs://...date=2015-11-10/000002_0:0+387632]
!= [
org.apache.hadoop.hive.ql.io.orc.OrcInputFormat:hdfs://...date=2015-11-10/000002_0:3+387427]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
