TabuLynx

TabuLynx is a local desktop spreadsheet-style viewer for tabular files. It is meant for opening, exploring, filtering, sorting, querying, and exporting local datasets without starting a web server or using a full spreadsheet suite.

It supports:

• CSV
• XLSX
• Parquet

The core is built on polars, and the desktop interface is built with PySide6.

Features

• Native desktop table viewer for local tabular files
• Data-type aware filters for text, numbers, booleans, dates, and datetimes
• Multi-filter and multi-sort support
• Global search across columns
• Column visibility controls
• Explicit pagination with configurable page size
• Row details side panel
• Column statistics side panel
• Dataset summary side panel
• Export of the current view to CSV or Parquet
• SQL queries against the current view using Polars SQL
• Temporary SQL result tabs
• SQL examples, SQL clear action, and local SQL history
• Query presets
• Multiple files in tabs
• Recent files and pinned files
• Last-session restore
• Query history per tab with back/forward navigation
• Context menu actions on cells and headers
• Optimized XLSX browsing through cached Parquet conversion
• Keyboard shortcuts for common actions

Quick Start

Install dependencies:

uv sync

Run the desktop app:

uv run tabulynx

Open a file directly:

uv run tabulynx open ./your-file.csv

You can also open:

uv run tabulynx open ./your-file.xlsx
uv run tabulynx open ./your-file.parquet

Basic Workflow

1. Open a csv, xlsx, or parquet file.
2. Use the global search box to quickly reduce the visible rows.
3. Add filters and sorts from the query area.
4. Use the Columns menu to show or hide columns.
5. Select a row to inspect full row values in the side panel.
6. Click a column header to view column statistics.
7. Use Export to save the current view as CSV or Parquet.

The current view means the result after applying search, filters, sorts, and visible columns.

SQL Support

The SQL panel uses Polars SQL and runs against the current tab's current view. The available table name is always current.

Example:

SELECT customer, revenue
FROM current
WHERE revenue > 100
ORDER BY revenue DESC
LIMIT 50

Supported today:

• Querying the current tab as current
• SELECT, WHERE, ORDER BY, LIMIT, and common Polars SQL expressions supported by Polars
• Aggregations and grouping supported by Polars SQL
• SQL over the current view, so active search, filters, sorts, and visible columns are applied before the SQL query runs
• Opening SQL results as temporary result tabs
• Quick SQL examples
• Local SQL history

Not supported yet:

• Cross-tab SQL queries or joins between multiple open files
• Persistent SQL result tabs across app restarts
• SQL editing of source files
• Database features such as indexes, transactions, stored procedures, or user-defined SQL functions
• Full DuckDB/PostgreSQL-style SQL compatibility

XLSX Handling

Excel files are handled differently from CSV and Parquet.

For xlsx, TabuLynx:

• Discovers sheet names up front
• Lets you switch sheets in the desktop UI
• Opens workbooks with a read-only ingestion path
• Converts the selected sheet into cached Parquet
• Uses the cached Parquet file for faster browsing, filtering, sorting, SQL, and pagination

This keeps the UI experience consistent while avoiding repeated direct Excel reads during exploration.

Important notes:

• The first open of a large sheet may take longer while the cache is created.
• The cache is invalidated when the source file metadata changes.
• Excel formulas are read as stored values when available.
• TabuLynx does not recalculate Excel formulas.
• TabuLynx does not preserve Excel formatting, formulas, merged cells, charts, or styles during export.

Keyboard Shortcuts

• Ctrl+O: open file
• Ctrl+F: focus global search
• Ctrl+Enter: run SQL in the current tab
• Esc: clear filters in the current tab
• Double-click a column header divider: autosize the column

Context Menus

Right-click a table cell for quick actions such as:

• Copy cell
• Copy row
• Filter by selected value
• Search/filter by selected text
• Hide the clicked column

Right-click a column header for quick column actions such as:

• Search in that column
• Clear filters for that column
• Sort ascending or descending
• Hide the column

Presets, Recent Files, And Session State

TabuLynx stores local app state under ~/.tabulynx/.

Current local state includes:

• Query presets
• SQL history
• Recent files
• Pinned files
• Last-session restore data
• XLSX Parquet cache

The app state is local to your machine and is not synced or uploaded by TabuLynx.

Library Usage

You can also use the core library directly:

from tabulynx import open_dataset

dataset = open_dataset("your-file.xlsx", sheet="Sheet1")

print(dataset.columns())
print(dataset.row_count())
print(dataset.page(0, 100))

Apply filters and sorts:

from tabulynx.core import ColumnFilter, SortRule, open_dataset

dataset = open_dataset("your-file.parquet")
dataset.add_filter(ColumnFilter(column="revenue", operator="gte", value="100"))
dataset.set_sorts([SortRule(column="revenue", descending=True)])

page = dataset.page(offset=0, limit=50)
print(page.rows)

Run SQL from Python:

dataset = open_dataset("your-file.csv")

result = dataset.sql_query(
    """
    SELECT customer, SUM(revenue) AS total_revenue
    FROM current
    GROUP BY customer
    ORDER BY total_revenue DESC
    """
)

print(result)

Sample Data

Sample data for local testing lives in:

• sample_data/filters_demo.csv
• sample_data/filters_demo.parquet

Regenerate the sample files:

uv run python scripts/create_sample_data.py

Generate a larger sample:

uv run python scripts/create_sample_data.py --rows 20000

Run Tests

Install development dependencies:

uv sync --group dev

Run the test suite:

uv run pytest -q

Project Structure

src/tabulynx/
  __init__.py
  __main__.py
  cli.py
  core.py
  qt.py
  viewer.py

sample_data/
  filters_demo.csv
  filters_demo.parquet

scripts/
  create_sample_data.py

tests/
  test_core.py

Current Limitations

• TabuLynx is a viewer and exploration tool, not a full spreadsheet editor.
• It does not currently edit cells or save changes back into the original source file.
• XLSX support uses cached Parquet conversion per selected sheet; very large or irregular workbooks may take time on first open.
• Excel formulas are not recalculated by TabuLynx.
• Excel formatting, styles, merged cells, charts, and workbook-specific layout features are not preserved.
• SQL currently runs only against the current tab's current view as the table current.
• SQL result tabs are temporary and are not restored in the next session.
• Cross-tab SQL joins are not supported yet.
• Extremely large datasets may still depend on available memory and local disk cache.
• The UI is focused on desktop use and is not intended for browser or mobile workflows.

Roadmap Ideas

• Cross-tab SQL queries
• Better SQL editor experience
• More test coverage for UI state and Excel edge cases
• More robust XLSX type inference for irregular workbooks
• Optional app packaging with PyInstaller or Nuitka
• Screenshots and release packaging documentation