Principle:Apache Druid Publication Settings

Knowledge Sources	Apache Druid Druid Datasource
Domains	Data_Ingestion, Data_Management
Last Updated	2026-02-10 00:00 GMT

Overview

A data publication principle that configures the target datasource name and write mode (append vs. replace) for ingested data.

Description

Publication Settings determine where ingested data lands in Druid and how it interacts with existing data:

• Datasource name: The target table name where segments are published. This is the name used in SQL queries (SELECT * FROM datasource_name).
• Append mode: Whether new segments are added alongside existing segments (append) or replace segments in the same time intervals (overwrite).

The datasource name must be unique within the Druid cluster. Appending is useful for incremental ingestion, while overwriting is used for full re-ingestion or corrections.

Usage

Use this principle as the final configuration step before spec review and submission. The datasource name and append mode determine how the ingestion interacts with the Druid metadata store and existing segments.

Theoretical Basis

Publication follows a write mode decision model:

Publication = {
  dataSource: string,           // Target datasource name
  appendToExisting: boolean     // true = append, false = overwrite
}

Append mode:
  New segments added alongside existing segments for the same time intervals

Overwrite mode:
  New segments replace existing segments in the same time intervals
  Atomic swap occurs after all new segments are built

Related Pages

Implemented By

• Implementation:Apache_Druid_Publication_Config_Form