[jira] [Updated] (SPARK-13212) Provide a way to unregister data sources from a SQLContext

["Hyukjin Kwon (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/SPARK-13212?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Hyukjin Kwon updated SPARK-13212:
---------------------------------
    Labels: bulk-closed  (was: )

> Provide a way to unregister data sources from a SQLContext
> ----------------------------------------------------------
>
>                 Key: SPARK-13212
>                 URL: https://issues.apache.org/jira/browse/SPARK-13212
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 1.6.0
>            Reporter: Daniel Darabos
>            Priority: Major
>              Labels: bulk-closed
>
> We allow our users to run SQL queries on their data via a web interface. We create an isolated SQLContext with {{sqlContext.newSession()}}, create their DataFrames in this context, register them with {{registerTempTable}}, then execute the query with {{isolatedContext.sql(query)}}.
> The issue is that they have the full power of Spark SQL at their disposal. They can run {{SELECT * FROM csv.`/etc/passwd`}}. This specific syntax can be disabled by setting {{spark.sql.runSQLOnFiles}} (a private, undocumented configuration) to {{false}}. But creating a temporary table (http://spark.apache.org/docs/latest/sql-programming-guide.html#loading-data-programmatically) would still work, if we had a HiveContext.
> As long as all DataSources on the classpath are readily available, I don't think we can be reassured about the security implications. So I think a nice solution would be to make the list of available DataSources a property of the SQLContext. Then for the isolated SQLContext we could simply remove all DataSources. This would allow more fine-grained use cases too.
> What do you think?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org