Description
Our customers and support team continuously face to the situation when setting a config parameter via spark.conf.set() does not may any effects. It is not clear from parameter's name is it static parameter or one of the parameter that can be set at runtime for current session state. It would be useful to have a method of RuntimeConfig which could tell to an user - does the given parameter may effect on the current behavior if he/she change it in the spark-shell or running notebook. The method can have the following signature:
def isModifiable(key: String): Boolean
Any config parameter can be checked by using the syntax like this:
scala> spark.conf.isModifiable("spark.sql.sources.schemaStringLengthThreshold") res0: Boolean = false
or for Spark Core parameter:
scala> spark.conf.isModifiable("spark.task.cpus") res1: Boolean = false
Attachments
Issue Links
- is related to
-
SPARK-33434 Document spark.conf.isModifiable()
- Resolved
- links to