[
https://issues.apache.org/jira/browse/SPARK-21054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Wang Haihua updated SPARK-21054:
--------------------------------
Description:
When we transfer our SQL in production from Hive to SparkSQL, we found some
SQL like {{reset dfs.blocksize;}} will throw error in SparkSQL.
We found cause is,
after SPARK-15330, Spark 2.0.0 support Reset command wich will Reset all
properties.
But after Hive 2.1.1 (HIVE-14418), Hive also support Reset Command for specific
property.
SQL compatibility with Hive will ease user more, lower the cost that transfer
SQL from Hive to SparkSQL.
was:
When we transfer our SQL in production from Hive to SparkSQL, we found some
SQL like {{reset dfs.blocksize;}} will throw error in SparkSQL.
We found cause is,
after SPARK-15330, Spark 2.0.0 support Reset command wich will Reset all
properties.
But after Hive 2.1.1 (HIVE-14418), Hive also support Reset Command for specific
property.
SQL compatibility with Hive will ease user more, lower the cost that transfer
SQL from Hive to SparkSQL.
> Reset Command support reset specific property which is compatible with Hive
> ---------------------------------------------------------------------------
>
> Key: SPARK-21054
> URL: https://issues.apache.org/jira/browse/SPARK-21054
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 2.1.1
> Reporter: Wang Haihua
> Priority: Minor
> Fix For: 2.1.1
>
>
> When we transfer our SQL in production from Hive to SparkSQL, we found some
> SQL like {{reset dfs.blocksize;}} will throw error in SparkSQL.
> We found cause is,
> after SPARK-15330, Spark 2.0.0 support Reset command wich will Reset all
> properties.
> But after Hive 2.1.1 (HIVE-14418), Hive also support Reset Command for
> specific property.
> SQL compatibility with Hive will ease user more, lower the cost that transfer
> SQL from Hive to SparkSQL.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
