A CSV/TSV plugin for Tabularis, the lightweight database management tool.
This plugin turns a folder of .csv or .tsv files into a queryable database. Each file becomes a table. Full SQL support via an in-memory SQLite engine — JOIN across files, GROUP BY, window functions, CTEs.
Zero dependencies — pure Python 3.8+ standard library. No pip install.
Discord - Join our discord server and chat with the maintainers.
- Any folder, instantly — point Tabularis at any directory and start querying.
- Auto-delimiter detection — automatically detects
,;\t|separators viacsv.Sniffer. - Full SQL — uses SQLite as the query engine:
JOIN,GROUP BY, subqueries, CTEs, window functions. - Schema Inspection — browse tables and columns in the sidebar explorer.
- ER Diagram — visualize relationships across your CSV files.
- SQL Execution — run any query with automatic pagination.
- Zero dependencies — ships as a single Python file with no external packages.
- Cross-platform — works on Linux, macOS, and Windows wherever Python 3.8+ is installed.
Open Settings → Available Plugins in Tabularis and install CSV Folder from the plugin registry.
- Download the latest
csv-plugin.zipfrom the Releases page. - Extract the archive.
- Copy
plugin.pyandmanifest.jsoninto the Tabularis plugins directory:
| OS | Plugins Directory |
|---|---|
| Linux | ~/.local/share/tabularis/plugins/csv/ |
| macOS | ~/Library/Application Support/com.debba.tabularis/plugins/csv/ |
| Windows | %APPDATA%\com.debba.tabularis\plugins\csv\ |
- Make the plugin executable (Linux/macOS):
chmod +x ~/.local/share/tabularis/plugins/csv/plugin.py- Restart Tabularis.
Python 3.8 or newer must be available as python3 in your PATH.
The plugin is a single Python script that communicates with Tabularis through JSON-RPC 2.0 over stdio:
- Tabularis spawns
plugin.pyas a child process. - Requests are sent as newline-delimited JSON-RPC messages to the plugin's
stdin. - Responses are written to
stdoutin the same format.
On first connection, the plugin scans the target folder, loads every .csv / .tsv file into an in-memory SQLite database, and keeps it alive for the entire session. SQLite handles all query execution.
All debug output is written to stderr and appears in Tabularis's log viewer — stdout is reserved exclusively for JSON-RPC responses.
| Method | Description |
|---|---|
test_connection |
Verify folder exists and contains CSV/TSV files |
get_databases |
Returns the folder name as the database name |
get_tables |
List all CSV/TSV files as tables |
get_columns |
Get column names for a table |
execute_query |
Execute SQL with pagination support |
get_schema_snapshot |
Full schema dump in one call (used for ER diagrams) |
get_all_columns_batch |
All columns for all tables in one call |
get_all_foreign_keys_batch |
Returns empty (no FK constraints in CSV) |
Test the plugin directly from your shell without opening Tabularis:
chmod +x plugin.py
echo '{"jsonrpc":"2.0","method":"test_connection","params":{"params":{"driver":"csv","database":"/path/to/example","host":null,"port":null,"username":null,"password":null,"ssl_mode":null}},"id":1}' \
| python3 plugin.pyA convenience script is provided to copy the plugin directly into your Tabularis plugins folder:
./sync.shThe example/ folder contains three related CSV files (users, orders, products):
./sync.sh
# In Tabularis, connect to the absolute path of the example/ folderThen try in the SQL editor:
SELECT u.name, SUM(CAST(p.price AS REAL) * CAST(o.quantity AS INTEGER)) AS total_spent
FROM orders o
JOIN users u ON o.user_id = u.id
JOIN products p ON o.product_id = p.id
GROUP BY u.name
ORDER BY total_spent DESC;- Language: Python 3.8+ (standard library only)
- Query engine: SQLite (via
sqlite3stdlib module) - Protocol: JSON-RPC 2.0 over stdio
Apache License 2.0