Connecting to DuckLake
The ways Coco Alemana can connect to a DuckLake — from a local file to a remote catalog backed by object storage.
How DuckLake is Put Together
Section titled “How DuckLake is Put Together”A DuckLake keeps two things separate:
- A catalog — the metadata that tracks tables, schemas, and snapshots.
- Data files — the actual rows, stored as Parquet.
Coco Alemana connects to both, and supports a specific set of options for each.
Supported Catalog Types
Section titled “Supported Catalog Types”The catalog is the database that stores the DuckLake metadata. Coco Alemana supports:
- DuckLake — a DuckLake catalog file (a
.ducklakefile). - Postgres — a PostgreSQL database.
Supported Data Locations
Section titled “Supported Data Locations”The data files hold the actual rows. Coco Alemana supports:
- Local — data files on your Mac.
- Amazon S3 — data files in an S3 bucket.
Local DuckLake
Section titled “Local DuckLake”The simplest setup uses a DuckLake catalog with Local data files, so the whole DuckLake lives on your Mac as a single .ducklake file. No account is required — you just open the file.
- Open the
.ducklakefile from Finder, or through Coco Alemana’s load-data sheet. - Choose the table you want to explore.
Remote DuckLake
Section titled “Remote DuckLake”A remote DuckLake keeps its catalog in a database and its data files in object storage. In practice this means a Postgres catalog with Amazon S3 data files. Connect each one once, then add the DuckLake.
Connect the Catalog
Section titled “Connect the Catalog”Add the Postgres account under External Accounts.... See Connect to Postgres.
Connect the Data Storage
Section titled “Connect the Data Storage”Add the Amazon S3 account the same way. See Connect to AWS.
Add the DuckLake
Section titled “Add the DuckLake”Once the catalog and storage are connected, add the DuckLake itself. Coco Alemana lists its tables and snapshots so you can start exploring.
