DuckDB as LakeXpress DB

Like SQLite, DuckDB needs no server: the LakeXpress DB is a single file. DuckDB can only be used as the LakeXpress DB, not as a source database.

1. Add it to the credentials file

There is nothing to create beforehand: LakeXpress creates the file, and any missing folder in its path, on first use.

{
  "lxdb_duckdb": {
    "ds_type": "duckdb",
    "auth_mode": "filesystem",
    "info": {
      "filepath": "/var/lib/lakexpress/lakexpress_log.duckdb"
    }
  }
}

Parameter	Required	Description
ds_type	Yes	Must be "duckdb"
auth_mode	Yes	Must be "filesystem"
filepath	Yes	Path to the DuckDB file. In-memory mode (:memory:) is not supported: the run history must persist between runs.

Use an absolute path. A relative path is resolved against the directory LakeXpress is run from, so running it from another directory would create a new, empty LakeXpress DB there.

2. Create the LakeXpress tables

./LakeXpress lxdb init -a credentials.json --lxdb_auth_id lxdb_duckdb

When to use DuckDB

Good for: trying LakeXpress, development, single-user setups without a database server.

Avoid for: several users or machines sharing configurations, syncs running in parallel.

Troubleshooting

Unable to open database file: The user running LakeXpress cannot write to the file or to its folder. Check the permissions of the path in filepath, and the free disk space.

Database is locked: Another process has the file open, for example another LakeXpress run or a DuckDB client. Close it and run the command again.