Databay dbt example

A small dbt project that Databay builds in its ClickHouse warehouse from this repository. It reads five tables a Databay pipeline lands (users, query_logs, conversations, messages, clicks) and builds:

Folder	Models	Stored as
models/base	base_users, base_searches, base_conversations, base_messages, base_clicks	views
models/marts	dim_users, fct_search_sessions, rpt_weekly_active_users	tables, rebuilt daily (rpt_weekly_active_users hourly)

How Databay reads it

• Sources are the warehouse's tables, named by the warehouse name you see in Databay: source('warehouse', 'users'). Databay writes sources.yml itself, so this repository has none.
• databay.latest_rows(...) is a Databay macro that keeps the current version of each replicated row.
• dbt_project.yml may set only vars and folder configs (materialized, unique_key, tags, enabled, meta: {databay: {refresh: ...}}). Seeds, snapshots and packages are refused.
• Refresh comes from meta.databay.refresh: every 15m, every 1h, every 1d, after_sources or manual.

Connect it

In Databay, open a project's Pipelines canvas, click Git, choose GitHub, install the Databay app on this repository, then pick branch main and folder /. Every push to main syncs; every pull request gets a "Databay models" check.

Because it reads Databay's own warehouse tables and macros, this project builds inside Databay rather than with a local dbt run.