# VisiData — Terminal Spreadsheet Multitool for Data Exploration

> An interactive terminal application for exploring, transforming, and analysing tabular data from CSV, JSON, SQLite, Postgres, and dozens of other formats.

## Install

Save as a script file and run:

# VisiData — Terminal Spreadsheet Multitool for Data Exploration

## Quick Use
```bash
pip install visidata
# Open a CSV file
vd data.csv
# Open a SQLite database
vd database.sqlite
# Pipe data from another command
curl -s https://api.example.com/data.json | vd -f json
```

## Introduction
VisiData is a terminal-based tool for rapid exploration and manipulation of tabular data. It opens virtually any structured data format instantly and provides a keyboard-driven interface for sorting, filtering, aggregating, and joining datasets without writing code.

## What VisiData Does
- Opens CSV, TSV, JSON, SQLite, Postgres, Excel, Parquet, HDF5, and 30+ other formats
- Provides instant frequency analysis, histogram, and pivot table generation
- Supports regex search, column transformations, and Python expressions inline
- Saves modified data back to any supported format
- Handles datasets with millions of rows via lazy loading and streaming

## Architecture Overview
VisiData is a pure-Python application built on the curses library. Each data source is loaded through a format-specific loader plugin that yields rows lazily. The sheet abstraction manages columns, types, and sort state. A command system maps single keystrokes to operations, and users can extend functionality with Python snippets in a `.visidatarc` file.

## Self-Hosting & Configuration
- Install via pip, pipx, conda, or your distro's package manager
- No external dependencies for CSV/JSON; install optional packages for databases (psycopg2) or Parquet (pyarrow)
- Customise keybindings and default options in `~/.visidatarc`
- Set `options.color_default` and theme options for terminal colour schemes
- Use `--play` to replay saved command logs for reproducible analyses

## Key Features
- Opens any data format with a single command and no schema definition
- Keyboard-driven workflow is faster than GUI tools for exploratory analysis
- Frequency sheets and pivot tables are one keystroke away
- Command logging enables fully reproducible data exploration sessions
- Extensible via Python plugins and inline expressions

## Comparison with Similar Tools
- **pandas** — requires writing Python code; VisiData is interactive and visual
- **csvkit** — focused on CSV command-line operations; VisiData adds an interactive TUI
- **Miller** — streaming record processor; VisiData adds spreadsheet-style navigation
- **q** — SQL on CSV files; VisiData supports SQL and many more formats interactively
- **Tad** — GUI CSV viewer; VisiData runs in any terminal without a display server

## FAQ
**Q: Can it handle very large files?**
A: Yes. VisiData streams rows lazily and can handle multi-gigabyte files if sufficient memory is available for the visible window.

**Q: Is it scriptable?**
A: Yes. Command logs can be saved and replayed, and `.visidatarc` supports arbitrary Python.

**Q: Does it modify the source file?**
A: No. Changes exist in memory until you explicitly save to a file.

**Q: What Python version is required?**
A: Python 3.8 or later.

## Sources
- https://github.com/saulpw/visidata
- https://www.visidata.org

---
Source: https://tokrepo.com/en/workflows/asset-9b3d16bb
Author: Script Depot