This message was deleted.

First, to clarify, Druid segments cannot be "updated" ... they are immutable once created and can only be overshadowed, replaced, or deleted. Batch ingestion supports both append and overwrite modes. But to my knowledge streaming ingestion only supports append mode due to it's architecture of involving multiple ingestion tasks operating in parallel, each creating their own separate segments within the same time chunk, and the ability to ingest "late arrival" data which belongs to older time chunks ... there is no notion in this architecture to facilitate some sort of group-level overwriting of existing data.

You can always setup a restatement batch job to “backfill” older data. We do that a lot.

@John Kowtko @Didip Kerabat great!! thanks for the advice and clarification

some detail here on restatement ("overwrite"): https://druid.apache.org/docs/latest/data-management/update.html#overwrite