josephglanville commented on issue #6004: Automatically Determine Ingestion Schema URL: https://github.com/apache/incubator-druid/issues/6004#issuecomment-411498697 @doc987 InputRowParsers are actually specific code that is loaded as a Druid extension so rather than running externally of Druid and transforming the data into a format Druid already knows how to parse (like JSON) they can read the format directly and turn it into Druid's internal "InputRow" representation directly. This obviates the need for a configured schema in cases where the message contains the schema itself and there is a logic mapping of the source format to dimensions. The important part of the picture is that dimensionSpec/metricsSpec in the provided ingestion spec doesn't need to be respected by your InputRowParser, the segment indexing code only cares about values returned from the parser itself which encode this information along with the row contents itself. See the constructor for MapBasedInputRow for clarity.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
