You are viewing the RapidMiner Radoop documentation for version 7.6 - Check here for latest version

What’s New in RapidMiner Radoop 7.4.1?

Enhancements and bug fixes

The following improvements are part of RapidMiner Radoop 7.4.1.

A different database for UDFs can now be defined with new Radoop connection parameters (check Use custom database for UDFs)
Spark RM now applies log level setting of Studio / Server; added log limit parameter to limit the number of recorded log entries
Connection accesswhitelist on Server now supports groups besides users
An error is displayed before a new line character would mess up data in text file format (Apache Hive may not parse it well)
Spark Script now turns up when searching for PySpark and SparkR terms
MapReduce is no longer the default execution engine on Hadoop Data View / Execute query...
Empty annotations are no longer saved as table comments
Added more details to the Error during preprocessing input error message in SparkRM

BUGFIX: SparkRM no longer fails with memory issues due to too many log entries recorded by the subprocess
BUGFIX: Spark learners no longer fail with The Spark job failed. error when there are too many attributes (few thousands) in input
BUGFIX: SparkRM no longer throws The cluster can't provide [...] memory error, but uses the maximum container memory setting of the cluster
BUGFIX: Read CSV no longer fails with division by zero in case of small input file
BUGFIX: Fixed that Store in Hive failed with stating that table already exists when dropfirst was unchecked.
BUGFIX: Fixed SparkRM and Single Process Pushdown error (com/rapidminer/external/alphanum/AlphanumComparator$AlphanumCaseSensitivity) due to using different Studio versions on a cluster
BUGFIX: Fixed rare NullPointerException during SparkRM and Single Process Pushdown log collection
BUGFIX: Fixed invalid warnings of connection tests stating local Hadoop settings differ to remote when setting contained equal sign