Enable a mode where all Plans are executed from the Druid Index
hbutani opened this issue · comments
The usecase here is to use the DruidDataSource as a traditional Spark DataSource. Bring in all the data from Druid and execute the plan in Spark. Even in this mode, we will still try to optimize the Plan to push Joins/Aggregates to Druid.
Use the nonAggregateQueryHandling option, see SelectQueryTest for examples