This message was deleted.

Hi Vinaya, Druid leave some things "less" compressed in the segment, especially when it is loaded on the Historical, for faster retrieval. When ingested, the JSON is split into its elements, and each element is stored in dictionary encoded column format for fast retrieval. Dictionaries may not be encoded, and in recent releases there is a front-coding compression that is available. The original JSON is currently also stored in it's full form for some query retrieval purposes. My understanding was that it is stored in json-smile format and then compressed (e.g. lz4) ... if you are seeing something different please let us know. Thanks. John

Hi John, Thank you for the response. We checked its compressed as you have mentioned. We are setting "useSchemaDiscovery": true, is there any way to tell not to index a JSON Column and to just store as a Json string?