Bruin CLI

Appearance

Bruin — Iceberg on your laptop ​

Loads two tables from Frankfurter, a public exchange-rate API, into Apache Iceberg tables on your own disk. No cloud account, no credentials, no services to start.

  • raw/currencies.asset.yml — the list of currency codes and names.
  • raw/exchange_rates.asset.yml — the latest rates.

Run it ​

bash
bruin run iceberg-sqlite-local

That is the whole setup. The connection in .bruin.yml needs nothing filled in:

yaml
      iceberg:
        - name: "iceberg-default"
          catalog:
            type: sqlite
            path: "catalog.db"          # relative to .bruin.yml
          storage:
            type: local
            path: "warehouse"           # relative to .bruin.yml

Afterwards, catalog.db holds the table metadata and warehouse/ the data:

warehouse/raw.db/currencies/data/00000-0-….parquet
warehouse/raw.db/currencies/metadata/00001-….metadata.json

Read them back with DuckDB:

sql
SET unsafe_enable_version_guessing = true;
SELECT * FROM iceberg_scan('warehouse/raw.db/currencies');

Notes ​

An Iceberg connection is a catalog (where table metadata lives) and storage (where the data files go). This template picks the simplest of each: a SQLite file and a local directory. Everything else — Glue, REST, Postgres, Hive, and S3 or GCS storage — swaps in by changing those two blocks only. The assets never change.

Not for production — the point is to see Iceberg work before committing to infrastructure. See the Iceberg docs for the real backends.