Release 0.4.0

Released October 24th, 2016

Maven

<dependency>
  <groupId>com.sparklinedata</groupId>
  <artifactId>spl-accel_2.10</artifactId>
  <version>0.4.0</version>
</dependency>

or

<dependency>
  <groupId>com.sparklinedata</groupId>
  <artifactId>spl-accel-onesixone_2.10</artifactId>
  <version>0.4.0</version>
</dependency>

Downloads

• Sparkline Accelerator Spark 1.6.2 jar
• Sparkline Accelerator Spark 1.6.1 jar
• Source Code

ChangeLog

• Changes

Feature/Bug Fixes

• Support qualified tableNames; StarSchema tables and Druid DataSource can be defined across any number of databases.
• allow use of Druid's v2 GroupBy Engine. (Commit)
• rework SelectQuery Iterator for Druid 0.9.1 (Commit)
• Handle Task Cancellation before Query is issued (Commit)
• Choose between lonSum and Count Aggregator for counts (Commit)
• More JS support: DateFormat, IF functions (Commit, Commit)
• Support ThriftServer HA & LoadBalancing using Zookeeper(SPARK-11100) (Commit)
• update CostModel: set cardinality, numRows; special cost models for each query type (Commit)
• Support JS Average (Commit)
• Spark Shim Layering, support for Spark1.6.x
• Introduce SparklineData Modules (Commit, Commit)
• Support Logical Optimization: Pull up VCols into Agg ([Commit] (https://github.com/SparklineData/spark-druid-olap/commit/2d73cd2999c546e42356f5cea41970aa62ef8207))
• Introduce Druid Test Cluster env ([Commit] (https://github.com/SparklineData/spark-druid-olap/commit/e0a0d64fb386810d7993158e86d06b69638495c5)) and Travis CI integration.
• introduce query exts test for hyperlog, thetasketch and spatial queries (Commit...)
• Several Bug Fixes.