[
https://issues.apache.org/jira/browse/SPARK-19927?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15906809#comment-15906809
]
bruce xu commented on SPARK-19927:
----------------------------------
The way I found this bug was to move from hiveQL production scripts to spark
sql production environment.
My assume is that the Spark sql user behavior inherits the user behavior of
Hive and the majority of spark user may have this assumption I guess.
> SparkThriftServer2 can not get ''--hivevar" variables in spark 2.1
> ------------------------------------------------------------------
>
> Key: SPARK-19927
> URL: https://issues.apache.org/jira/browse/SPARK-19927
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 2.1.0
> Environment: CentOS 6.5,spark 2.1 build with mvn -Pyarn -Phadoop-2.6
> -Dhadoop.version=2.6.0 -Phive -Phive-thriftserver -Dscala-2.11
> Reporter: bruce xu
>
> suppose the content of test1.sql:
> -------------------------------------------------
> USE ${hivevar:db_name};
> -------------------------------------------------
>
> when execute: bin/spark-sql -f /tmp/test.sql --hivevar db_name=offline
> the output is:
> Error: org.apache.spark.sql.catalyst.parser.ParseException:
> no viable alternative at input '<EOF>'(line 1, pos 4)
> == SQL ==
> use
> ----^^^ (state=,code=0)
> so hivevar can not be read from CLI.
> the bug still appears with beeline command: bin/beeline -f /tmp/test2.sql
> --hivevar db_name=offline with test2.sql:
> ----------------------------------------
> !connect jdbc:hive2://localhost:10000 test test
> USE ${hivevar:db_name};
> ----------------------------------------
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
