DatriseAI-first ETL

Harvest DuckDB

AI-first ETL from Harvest into DuckDB. Governed entities, incremental sync, typed landing tables.

How Datrise loads Harvest into DuckDB

Datrise syncs Harvest's time entries, projects, clients, invoices, and utilization into DuckDB as a typed table per source entity in a DuckDB file. Flexible or custom fields land in JSON or STRUCT columns, and timestamps such as created, updated, and status changes are typed as TIMESTAMP WITH TIME ZONE.

Sync is incremental: Datrise uses rewrites changed entities into the local database (or Parquet) on each run, so re-runs update only what changed. Hive-partitioned Parquet by load date when exporting. DuckDB is single-writer and embedded, so Datrise produces a consistent file snapshot rather than concurrent streaming writes.

Ideal for local and notebook analytics without standing up a server.

Endpoints

Harvest: Time tracking and project profitability for services teams.

DuckDB: In-process analytics database for fast local OLAP.

How Harvest entities map to DuckDB

Harvest entityDuckDB objectNotes
time entriesharvest_time_entriesid PK · custom fields → JSON or STRUCT columns
projectsharvest_projectsid PK · linked to harvest_time_entries
clientsharvest_clientsid PK · linked to harvest_time_entries
invoicesharvest_invoicesid PK · linked to harvest_time_entries

FAQ

How does Datrise handle Harvest's custom fields in DuckDB?

Flexible values are stored as JSON or STRUCT columns, so new fields don't require a migration; strongly-typed fields — dates, numbers, and references — are promoted to native DuckDB types.

How does the Harvest to DuckDB sync stay up to date?

It runs incrementally — Datrise uses rewrites changed entities into the local database (or Parquet) on each run.

Related pipelines

Early access

Connect Harvest to DuckDB the easy way

Skip brittle scripts and manual exports. Join the waitlist to get a guided setup, AI-assisted mapping, and reliable incremental sync for this integration.