csvex

Explore, compare, inspect, and query tabular data from the terminal.

Open a file into a layered red interactive table, browse folders with live counts, compare related CSVs, inspect columns, export filtered views, run schema drift checks, and use SQL without leaving the shell.

What It Can Do

• open .csv, .tsv, .txt, compressed text files like .csv.gz, and .parquet
• launch a bundled guided demo with csvex demo
• launch a colorful TUI with csvex data.csv or browse folders with csvex explore
• fuzzy-find files and folders inside the explorer with /
• search rows, filter columns, sort data, pin columns, hide columns, and export visible views
• inspect a selected column with mini charts and run lightweight data quality checks
• keep favorites and recent files in the explorer
• compare two files and show added, removed, and changed rows
• summarize a folder of files and detect schema drift across it
• run SQL against a single file with an in-memory table named data
• read from stdin with csvex - ...

Quick Start

Editable install with global commands

./install-dev-global.sh

This creates .venv, installs the project in editable mode, and links both csvex and c into /usr/local/bin.

User install

python3 -m pip install --user .

Common Flows

Show the banner and quick-start guide

csvex
c

Launch the built-in demo

csvex demo
csvex demo --browse
c demo

csvex demo opens a bundled showcase dataset and starts with a guided overlay. From there you can press b to browse the rest of the demo pack, including a compare-friendly sales pair, a survey dataset, and a TSV inventory file.

Open a file in the interactive table view

csvex data.csv
c data.csv

Large text files automatically open in preview mode inside the TUI so startup stays responsive.

Browse a folder

csvex explore
csvex explore data/
c explore .

Compare two files

csvex compare before.csv after.csv --key id

Run SQL against one file

csvex sql data.csv "select country, count(*) as total from data group by country order by total desc"

Summarize a folder and check schema drift

csvex summary data/
csvex drift data/

Use the shell-friendly file commands

csvex data.csv head -n 20
csvex data.csv filter age '>' 30
csvex data.csv select name email salary
csvex data.csv inspect country
csvex data.csv quality
csvex data.csv export output.csv

Read from stdin

printf 'name\tage\nAna\t30\nBen\t41\n' | csvex - head -n 2

Commands

Command	Description
head [-n N]	Display the first N rows
filter <col> <op> <val>	Filter rows by value
select <cols...>	Show only selected columns
stats	Show numeric summary statistics
unique <col>	List unique values in a column
inspect <col>	Inspect one column in depth
quality	Run lightweight data quality checks
export <path>	Export the file to another path
explore [path]	Open the interactive explorer
demo [entry] [--browse]	Launch the bundled csvex showcase workspace
summary [path]	Summarize a folder or one file
drift [path]	Show schema drift across a folder
compare <left> <right> [--key col]	Compare two files
sql <file> "<query>"	Run an in-memory SQL query against one file

TUI Controls

Explorer

• Enter or l: open the selected folder or file
• h, u, or Backspace: go up one folder
• a: toggle folder browser and recursive all-files view
• F: toggle favorites view
• R: toggle recents view
• *: add or remove the selected item from favorites
• /: fuzzy-find files and folders in the current explorer view
• C: clear the explorer search filter
• m: show a folder summary overlay
• D: show schema drift for the current folder or root
• Up/Down or j/k: move selection
• PgUp/PgDn: page through entries
• g / G: jump to top or bottom
• r: rescan
• ?: help overlay
• q: quit

Viewer

• Up/Down or j/k: move the selected row
• Left/Right or h/l: scroll horizontally
• PgUp/PgDn: page through rows
• g / G: jump to top or bottom
• 0 / $: jump to the left or right edge
• [ / ]: move the active column
• /, n, N: search rows and cycle through matches
• f: filter the active column with a contains match
• C: clear search and filter state
• s / S: sort ascending or descending by the active column
• x / X: hide the active column or restore all columns
• p: pin or unpin the active column
• Enter: open the selected row detail view
• i: inspect the active column
• d: show file-wide data quality checks
• e: export the current visible view
• w: watch the file for changes and auto-reload it
• R: reload the active file immediately
• b: go back to the file list
• r: go back to files and rescan
• ?: help overlay
• q: quit

Supported Formats

• .csv
• .tsv
• .txt
• .csv.gz, .tsv.gz, and other pandas-supported compressed text variants
• .parquet
• stdin via -

Contributing

csvex is still intentionally compact enough to hack on quickly. Most of the behavior lives in csvex.py, so it is easy to trace loader logic, CLI dispatch, explorer state, and TUI rendering.

Good contribution areas:

• stronger large-file support beyond preview mode
• richer SQL features and multi-file joins
• better schema drift reports and validations
• smarter data quality checks
• deeper tests around curses behavior and parser edge cases

Basic development loop:

./install-dev-global.sh
python3 -m py_compile csvex.py
c
c test_data.csv
c explore .

License

MIT