[ https://issues.apache.org/jira/browse/HIVE-3454?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14347042#comment-14347042 ]
Aihua Xu commented on HIVE-3454: -------------------------------- Was not aware of LLAP. Seems reasonable to defer the issue since I guess LLAP needs to handle many other similar thread-safe problems like this. Correct me if I'm wrong with the hadoop execution: currently (without LLAP) each ExecMapper and ExecReducer runs in its own JVM and it could spin off new threads. Each ExecMapper and ExecReducer handles one job from a query and we want those threads to share that same static variable value. With LLAP, if ExecMapper runs in a thread and doesn't spin off new threads, then thread-local would work, but we still need to handle current existing scenario. If new threads are spun off, thread-local actually won't work. > Problem with CAST(BIGINT as TIMESTAMP) > -------------------------------------- > > Key: HIVE-3454 > URL: https://issues.apache.org/jira/browse/HIVE-3454 > Project: Hive > Issue Type: Bug > Components: Types, UDF > Affects Versions: 0.8.0, 0.8.1, 0.9.0, 0.10.0, 0.11.0, 0.12.0, 0.13.0, > 0.13.1 > Reporter: Ryan Harris > Assignee: Aihua Xu > Labels: newbie, newdev, patch > Attachments: HIVE-3454.1.patch.txt, HIVE-3454.2.patch, > HIVE-3454.3.patch, HIVE-3454.4.patch, HIVE-3454.patch > > > Ran into an issue while working with timestamp conversion. > CAST(unix_timestamp() as TIMESTAMP) should create a timestamp for the current > time from the BIGINT returned by unix_timestamp() > Instead, however, a 1970-01-16 timestamp is returned. -- This message was sent by Atlassian JIRA (v6.3.4#6332)