what kind of ELT patterns are people seeing for duckdb? Is there anything important to consider or do differently compared to old school data warehousing?
r
Ryan Boyd
12/07/2022, 10:15 PM
hi @Andy Clayton - great question. DuckDB itself is essentially a SQL engine. super easy to ingest parquet files and produce them. [with iceberg and more coming]. a lot of the usage today is embedded into applications or on the local desktop [sometimes for super cheap transformation work!], but some folks are starting to use it in cloud environments for more automated workloads and transformations — if you search twitter for duckdb and lambda, you’ll see some great content.
in terms of integration with other toolsets, josh wills built dbt-duckdb: https://github.com/jwills/dbt-duckdb. the dagster folks also wrote a good blog post on using duckdb with dagster: https://dagster.io/blog/duckdb-data-lake
there’s less in the airbyte community that i’ve found, but i’m sure it’s coming 🙂. Interestingly, they have a pretty good state of the world over here in their data glossary: https://glossary.airbyte.com/term/duckdb/.