xsv

xsv is a suite of fast CSV manipulation commands bundled into a single binary. Rather than reaching for awk, cut, or a Python script every time you want to explore a CSV file, xsv gives you a consistent, composable set of subcommands that pipe cleanly together.

Features

• Indexing — build an index for instant random access and parallel operations on large files
• Slicing and selecting — extract specific rows, columns, or ranges without loading the whole file
• Searching — filter rows by regex match on one or more columns
• Joining — perform inner, left, right, and full outer joins between two CSV files
• Sorting and deduplicating — sort by one or more columns, remove duplicate rows
• Statistics — compute count, min, max, mean, standard deviation, and median for each column
• Frequency tables — count distinct values per column
• Flattening — convert a CSV into a human-readable vertical key/value format
• Fast — built with performance in mind; handles files far larger than memory when indexed

Installation

cargo install xsv

Or via your package manager:

# Debian / Ubuntu
apt install xsv

# Fedora
dnf install xsv

# macOS
brew install xsv

# Arch Linux
yay -S xsv

Usage

# Show the first 5 rows with headers
xsv slice -l 5 data.csv | xsv table

# Select specific columns
xsv select name,age,email data.csv

# Search for rows where any field matches a pattern
xsv search "New York" data.csv

# Search in a specific column
xsv search -s city "New York" data.csv

# Print summary statistics for each column
xsv stats data.csv | xsv table

# Count distinct values in the "country" column
xsv frequency -s country data.csv

# Join two CSV files on a common column
xsv join id users.csv user_id orders.csv

# Sort by a column
xsv sort -s age data.csv

# Remove duplicate rows
xsv dedup data.csv

# Pretty-print as an aligned table (great for terminal viewing)
xsv table data.csv

# Convert to a vertical key/value view (useful for wide CSVs)
xsv flatten data.csv

# Count rows
xsv count data.csv

Indexing for large files

For large CSV files, building an index unlocks parallel processing and instant random access:

# Build an index
xsv index data.csv

# Now slice operations are instant, even on huge files
xsv slice -s 100000 -l 10 data.csv

# Stats run in parallel across cores
xsv stats --everything data.csv

Composing commands

xsv subcommands pipe naturally:

# Find all users from Germany, select name and email, pretty-print
xsv search -s country Germany users.csv \
  | xsv select name,email \
  | xsv table

See also

qsv is an actively maintained fork of xsv with many additional subcommands and ongoing bug fixes, worth considering if you need features beyond what xsv provides.