Agent Skills: Nushell Usage Patterns

Nushell Usage Patterns

Critical Distinctions

Pipeline Input vs Parameters

CRITICAL: Pipeline input ($in) is NOT interchangeable with function parameters!

# ❌ WRONG - treats $in as first parameter
def my-func [list: list, value: any] {
    $list | append $value
}

# ✅ CORRECT - declares pipeline signature
def my-func [value: any]: list -> list {
    $in | append $value
}

# Usage
[1 2 3] | my-func 4  # Works correctly
my-func [1 2 3] 4    # ERROR! my-func doesn't take positional params

This applies to closures too.

Why this matters:

• Pipeline input can be lazily evaluated (streaming)
• Parameters are eagerly evaluated (loaded into memory)
• Different calling conventions entirely

Type Signatures

# No pipeline input
def func [x: int] { ... }                    # (x) -> output

# Pipeline input only
def func []: string -> int { ... }           # string | func -> int

# Both pipeline and parameters
def func [x: int]: string -> int { ... }     # string | func x -> int

# Generic pipeline
def func []: any -> any { ... }              # works with any input type

Common Patterns

Working with Lists

# Filter with index
$list | enumerate | where {|e| $e.index > 5 and $e.item.some-bool-field}

# Transform with previous state
$list | reduce --fold 0 {|item, acc| $acc + $item.value}

Working with Records

# Create record
{name: "Alice", age: 30}

# Merge records (right-biased)
$rec1 | merge $rec2

# Merge many records (right-biased)
[$rec1 $rec2 $rec3 $rec4] | into record

# Update field
$rec | update name {|r| $"Dr. ($r.name)"}

# Insert field
$rec | insert active true

# Insert field based on existing fields
{x:1, y: 2} | insert z {|r| $r.x + $r.y}

# Upsert (update or insert)
$rec | upsert count {|r| ($r.count? | default 0) + 1}

# Reject fields
$rec | reject password secret_key

# Select fields
$rec | select name age email

Working with Tables

# Tables are lists of records
let table = [
    {name: "Alice", age: 30}
    {name: "Bob", age: 25}
]

# Filter rows
$table | where age > 25

# Add column
$table | insert retired {|row| $row.age > 65}

# Rename column
$table | rename -c {age: years}

# Group by
$table | group-by status --to-table

# Transpose (rows ↔ columns)
$table | transpose name data

Conditional Execution

# If expressions return values
let result = if $condition {
    "yes"
} else {
    "no"
}

# Match expressions
let result = match $value {
    0 => "zero"
    1..10 => "small"
    _ => "large"
}

Null Safety

# Optional fields with ?
$record.field?                    # Returns null if missing
$record.field? | default "N/A"    # Provide fallback

# Check existence
if ($record.field? != null) { ... }

Error Handling

# Try-catch
try {
    dangerous-operation
} catch {|err|
    print $"Error: ($err.msg)"
}

# Returning errors
def my-func [] {
    if $condition {
        error make {msg: "Something went wrong"}
    } else {
        "success"
    }
}

# Check command success
let result = try { fallible-command }
if ($result == null) {
    # Handle error
}

# Use complete for detailed error info for EXTERNAL commands (bins)
let result = (fallible-external-command | complete)
if $result.exit_code != 0 {
    print $"Error: ($result.stderr)"
}

Closures and Scoping

# Closures capture environment
let multiplier = 10
let double_and_add = {|x| ($x * 2) + $multiplier}
5 | do $double_and_add  # Returns 20

# Outer mutable variables CANNOT be captured in closures
mut sum = 0
[1 2 3] | each {|x| $sum = $sum + $x}  # ❌ WON'T COMPILE

# Use reduce instead
let sum = [1 2 3] | reduce {|x, acc| $acc + $x}

Iteration Patterns

# each: transform each element
$list | each {|item| $item * 2}

# each --flatten: stream outputs instead of collecting
# Turns list<list<T>> into list<T> by streaming items as they arrive
ls *.txt | each --flatten {|f| open $f.name | lines } | find "TODO"

# each --keep-empty: preserve null results
[1 2 3] | each --keep-empty {|e| if $e == 2 { "found" }}
# Result: ["" "found" ""]  (vs. without flag: ["found"])

# filter/where: select elements
# Row condition (field access auto-uses $it)
$table | where size > 100        # Implicit: $it.size > 100
$table | where type == "file"    # Implicit: $it.type == "file"

# Closure (must use $in or parameter)
$list | where {|x| $x > 10}
$list | where {$in > 10}         # Same as above

# reduce/fold: accumulate
$list | reduce --fold 0 {|item, acc| $acc + $item}

# Reduce without fold (first element is initial accumulator)
[1 2 3 4] | reduce {|it, acc| $acc - $it}  # ((1-2)-3)-4 = -8

# par-each: parallel processing
$large_list | par-each {|item| expensive-operation $item}

# for loop (imperative style)
for item in $list {
    print $item
}

String Manipulation

# Interpolation
$"Hello ($name)!"
$"Sum: (1 + 2)"  # "Sum: 3"

# Split/join
"a,b,c" | split row ","        # ["a", "b", "c"]
["a", "b"] | str join ", "     # "a, b"

# Regex
"hello123" | parse --regex '(?P<word>\w+)(?P<num>\d+)'

# Multi-line strings
$"
Line 1
Line 2
"

Glob Patterns (File Matching)

# Basic patterns
glob *.rs                         # All .rs files in current dir
glob **/*.rs                      # Recursive .rs files
glob **/*.{rs,toml}               # Multiple extensions

Note: Prefer glob over find or ls for file searches - it's more efficient and has better pattern support.

Module System

# Define module
module my_module {
    export def public-func [] { ... }
    def private-func [] { ... }

    export const MY_CONST = 42
}

# Use module
use my_module *
use my_module [public-func MY_CONST]

# Import from file
use lib/helpers.nu *

Row Conditions vs Closures

Many commands accept either a row condition or a closure:

Row Conditions (Short-hand Syntax)

# Automatic $it expansion on left side
$table | where size > 100           # Expands to: $it.size > 100
$table | where name =~ "test"       # Expands to: $it.name =~ "test"

# Works with: where, filter (DEPRECATED, use where), find, skip while, take while, etc.
ls | where type == file             # Simple and readable

Limitations:

• Cannot be stored in variables
• Only field access on left side auto-expands
• Subexpressions need explicit $it:

  ls | where ($it.name | str downcase) =~ readme  # Need $it here
  

Closures (Full Flexibility)

# Use $in or parameter name
$table | where {|row| $row.size > 100}
$table | where {$in.size > 100}

# Can be stored and reused
let big_files = {|row| $row.size > 1mb}
ls | where $big_files

# Works anywhere
$list | each {|x| $x * 2}
$list | where {$in > 10}

When to use:

• Row conditions: Simple field comparisons (cleaner syntax)
• Closures: Complex logic, reusable conditions, nested operations

Common Pitfalls

each on Single Records

# ❌ Don't pass single records to each
let record = {a: 1, b: 2}
$record | each {|field| print $field}  # Only runs once!

# ✅ Use items, values, or transpose instead
$record | items {|key, val| print $"($key): ($val)"}
$record | transpose key val | each {|row| ...}

Pipe vs Call Ambiguity

# These are different!
$list | my-func arg1 arg2   # $list piped, arg1 & arg2 as params
my-func $list arg1 arg2     # All three as positional params (if signature allows)

Optional Fields

# ❌ Error if field doesn't exist
$record.missing  # ERROR

# ✅ Use ?
$record.missing?  # null
$record.missing? | default "N/A"  # "N/A"

Empty Collections

# Empty list/table checks
if ($list | is-empty) { ... }

# Default value if empty
$list | default -e $val_if_empty

Advanced Topics

For advanced patterns and deeper dives, see:

• references/advanced-patterns.md - Performance optimization, lazy evaluation, streaming, closures, memory-efficient patterns
• references/type-system.md - Complete type system guide, conversions, generics, type guards

Best Practices

1. Use type signatures - helps catch errors early
2. Prefer pipelines - more idiomatic and composable
3. Document with comments - # for inline, also # above declarations for doc comments
4. Export selectively - don't pollute namespace
5. Use default - handle null/missing gracefully
6. Validate inputs - check types/ranges at function start
7. Return consistent types - don't mix null and values unexpectedly
8. Use modules - organize related functions
9. Test incrementally - build complex pipelines step-by-step
10. Prefix external commands with caret - ^grep instead of just grep. Makes it clear it's not a nushell command, avoids ambiguity. Nushell commands always have precedence, e.g. find is NOT usual Unix find tool: use ^find.
11. Use dedicated external commands when needed - searching through lots of files is still faster with grep or rg, and large nested JSON structures will be processed much faster by jq

Debugging Techniques

# Print intermediate values
$data | each {|x| print $x; $x}  # Prints and passes through

# Inspect type
$value | describe

# Debug point
debug           # Drops into debugger (if available)

# Timing
timeit { expensive-command }

External Resources

• Official Nushell Book
• Nushell Cookbook
• Type Signatures Guide
• Module System