Exploring Nushell: A Rust-Powered Cross-Platform Shell

What Makes Nushell Different

Nushell represents a fundamental shift in how developers interact with their command line. Unlike traditional shells that output plain text, Nushell embraces structured data from the ground up, allowing you to treat command output as queryable tables rather than strings to parse. Built entirely in Rust, it delivers cross-platform consistency while maintaining performance and memory safety.

For teams building modern web applications, adopting tools that emphasize structured data and type safety aligns well with modern web development practices that prioritize reliability and maintainability.

Key differentiators:

• Traditional shells output text; Nushell outputs structured data
• Rust-based implementation provides cross-platform consistency
• Commands return tables that can be filtered, sorted, and queried
• The Unix philosophy of composable commands enhanced with structured data

The Structured Data Advantage

The first thing you'll notice when running a command like ls is that instead of a block of text coming back, you get a structured table. This table does more than just format the output nicely. Like a spreadsheet, it allows you to work with the data interactively.

Installation and Cross-Platform Support

Nushell's Rust-based implementation ensures consistent behavior across all major operating systems. Whether you're on Windows, macOS, or Linux, you'll experience the same powerful features.

Installation Methods

macOS:

brew install nushell

Windows:

winget install Nushell.Nushell
# or
scoop install nushell

Linux:

cargo install nushell
# or via package manager
# Debian/Ubuntu: sudo apt install nushell
# Arch: sudo pacman -S nushell

Cross-Platform Commands

Nushell provides built-in versions of common system commands that behave identically across platforms:

• ls - Directory listing with structured output
• ps - Process information as tables
• date - Date and time operations
• rm, cp, mv - File operations with type safety

While Nushell provides cross-platform built-ins, you can still run external system commands by prefacing them with the caret sigil ^ when platform-specific behavior is needed.

For organizations managing diverse development environments, the consistency Nushell provides can significantly reduce the operational overhead of maintaining cross-platform automation workflows.

Core Command Patterns

Commands Output Data

When you run ls in Nushell, you don't get a block of text--you get a structured table with named columns:

ls
# => ╭────┬─────────────────────┬──────┬───────────┬──────────────╮
# => │ # │ name │ type │ size │ modified │
# => ├────┼─────────────────────┼──────┼───────────┼──────────────┤
# => │ 0 │ Cargo.toml │ file │ 9.2 KiB │ 15 hours ago │
# => │ 1 │ README.md │ file │ 12.0 KiB │ 15 hours ago │
# => ...

The Pipeline Philosophy

Traditional Unix pipelines connect text streams. Nushell pipelines connect data streams:

ls | sort-by size | reverse

This command sorts files by size in descending order. The sort-by command operates on the size column values, not text--and it understands that 1.1 KiB is larger than 812 bytes because the column has the filesize type.

Filtering with Where

Filter data using the where command:

ls | where size > 10kb

This returns only files larger than 10 kilobytes, with type-aware comparison.

Working with System Data

Process Monitoring

Nushell's ps command provides a cross-platform view of system processes as structured data:

ps
# => ╭───┬──────┬──────┬──────┬─────────┬──────┬──────────┬─────────╮
# => │ # │ pid │ ppid │ name │ status │ cpu │ mem │ virtual │
# => ├───┼──────┼──────┼──────┼─────────┼──────┼──────────┼─────────┤
# => │ 0 │ 1 │ 0 │ init │ Sleeping│ 0.00 │ 1.2 MiB │ 2.2 MiB │
# => ...

Filter for running processes:

ps | where status == "Running"

The Describe Command

Examine the types of data you're working with:

ps | describe
# => table<pid: int, ppid: int, name: string, status: string, cpu: float, mem: filesize, virtual: filesize> (stream)

This reveals that Nushell knows mem and virtual are filesize types, and cpu is a float.

Advanced Pipeline Techniques

The $in Variable

When a command needs an argument instead of pipeline input, use the $in variable:

ls | sort-by size | reverse | first | get name | cp $in ~

This copies the largest file to your home directory, using $in to pass the pipeline output (the filename) to cp.

Data Navigation

Use get and select to navigate structured data:

• get <column> - Retrieve a specific value from a record
• select <columns> - Choose specific columns from a table

ls | select name size | get name
# Returns just the names as a list

Exploring Nested Data

The explore command provides an interactive browser for any structured data:

help commands | explore

Navigate with arrow keys, press Enter to telescope into nested structures, and Esc to return.

Best Practices for Nushell

Command Design Principles

• Prefer pipeline input over positional arguments when possible
• Use multi-line commands for readability in complex pipelines
• Leverage type-aware operations instead of string parsing
• Create custom commands for frequently repeated operations

Writing Readable Scripts

# Break complex pipelines across multiple lines
ls
| where type == "file"
| where size > 1mb
| sort-by modified
| reverse
| first 10

# Define reusable commands
def large-files [] {
 ls | where type == "file" | where size > 10mb
}

Configuration

Customize Nushell through config.nu:

• Set up aliases for frequently used commands
• Define environment variables for your workflow
• Configure the prompt and display preferences
• Load custom modules and scripts

By adopting structured data workflows like those Nushell provides, teams can build more maintainable automation systems. Our AI automation services help organizations implement similar principles across their development and operational processes.

Sources

1. Nushell Book - Quick Tour
2. Nushell Official Documentation
3. LogRocket: Exploring Nushell
4. Nushell GitHub Repository