developing-incremental-models

What problem does it solve?

This Skill helps you develop and troubleshoot dbt incremental models so you get correct, up-to-date results without paying the cost of full refreshes or falling into common incremental pitfalls.

Core Features & Use Cases

• Strategy selection: Choose the right incremental strategy (table, incremental with merge/append/delete+insert, or insert_overwrite for partitioned warehouses) based on data patterns and performance constraints.
• Correctness guardrails: Define a truly unique unique_key, enforce a full-refresh baseline first, and validate incremental behavior to prevent data drift.
• Troubleshooting & optimization: Fix merge failures (duplicate keys), improve partition pruning, handle late-arriving data with lookback windows, and manage schema drift using on_schema_change.

Quick Start

Use this guidance to configure and run a new dbt model as incremental with a safe strategy, then verify it first with full-refresh and again with incremental runs.