DuckDBT: Not a database or a dbt adapter but a secret third thing
summary
Adapter glues DBT to DuckDB so DBT runs against local DuckDB instances and object stores without standing up a cloud warehouse; it supports Iceberg, SQLAlchemy backends, Excel plugin, and out-of-core hash aggregates for parquet-in-S3 workflows that replace Spark when data fits memory, and it appears faster and cheaper. A Python server called Buena Vista pretends to be Postgres so DBT talks to notebook kernels; column-level lineage uses DuckDB’s AST; the modern data stack in a box demo runs end-to-end and is production-tested.
AI-written summary · about
chapters (8)
- 0:00Start
- 0:59What, exactly, is dbt?
- 3:35Like Terraform For Analytical Data Models
- 4:30Data Modeling Without The Data Warehouse
- 6:15My Super Villain Origin Story
- 6:30March 2020
- 13:58The Modern Data Stack in a Box
- 17:23Buena Vista
related talks
43mUse dbt and DuckDB Instead of Databricks for Data Processing
30mDuckDB: Bringing analytical SQL directly to your Python shell
28mAnalyze data easily with duckdb - and the implications on data architectures
24mpg_duckdb: Ducking awesome analytics in Postgres
40mDuckDB beyond the notebook
39mPython and SQL: Better Together, Powered by DuckDB
05mDuckDB Powering Interactive Notebooks
27mIn Process Analytical Data Management with DuckDB