DuckDB
DuckDB files are found in container volumes by their first bytes, and on the host with the host scan. You can also add a file by its path.
| Plan | Pro and Team |
| Default ports | none (a file) |
What you can do
Section titled “What you can do”DuckDB database files, and Parquet files, open like SQLite files: a snapshot from a container, live from the host. CSV and JSON files can be added by path. Parquet, CSV and JSON files open as one read-only view named after the file (events.parquet → events.main.events). Browse, edit (live database files), the SQL editor with DuckDB’s SQL, Explain, exports and the ER diagram.
A least-privilege user
Section titled “A least-privilege user”Wharfbase works with any user; give it only what you want it to do.
Read only
Section titled “Read only”DuckDB has no users. A snapshot is always read-only, and so are Parquet, CSV and JSON files. For a live host file, Wharfbase needs read access to the file; it opens it read-only unless the connection allows writes. Read-only connections also run every statement in a read-only transaction.
Read and write
Section titled “Read and write”Writes to a live database file need write access to the file and its directory (DuckDB writes a .wal next to it). DuckDB locks a file it writes: while another program has it open for writing, Wharfbase can only open a snapshot of it.
Once its file is open, Wharfbase turns DuckDB’s file access off and locks its settings: SQL can’t read or write other files, attach databases, or install and load extensions. Each open file gets 1 GB of memory and 2 threads.