Tranquility Indexing

Hi ,

I want to send a stream of events from my application to druid that I can aggregate upon.

If I send the attributes in my events in form of a map (Key: attribute name and Value: Attribute value) and do that as Schemaless, is it reasonable to expect that I can query on an attribute name that was in the “key” of the map ?

case class SimpleEvent(ts: DateTime, fields: collection.mutable.Map[String, String]) {
  @JsonValue
  def toMap = fields ++ Map(DruidBeams.DefaultTimestampSpec.getTimestampColumn -> ts.toString())
}


DruidBeams.
.....

.rollup(DruidRollup(DruidDimensions.schemalessWithExclusions(util.Arrays.asList(“test”)), aggregators, QueryGranularity.MINUTE))


My event has 50attributes. Instead of defining a schema with 50 attributes , I am thinking of using the map with attribute name as key.

I see if I do that, I am unable to query on an attribute say "channel" that was one of the attribute key in Map. 

Thanks,
Deepika

Click here to Reply

Duplicated by https://groups.google.com/forum/#!topic/druid-development/0Vz98Q8hShI