Unicode Patterns¶

Ripgrep has excellent Unicode support enabled by default. All regex metacharacters are Unicode-aware.

Unicode Character Classes¶

Use \p{Property} syntax to match Unicode character properties:

# Find emoji
rg '\p{Emoji}'

# Find Greek text
rg '\p{Greek}+'

# Find alphabetic characters (all scripts)
rg '\p{Alphabetic}+'

# Find uppercase letters (all scripts)
rg '\p{Uppercase}+'

How Unicode Properties Work¶

Unicode properties categorize characters by script, category, or attribute. When ripgrep evaluates a \p{Property} pattern, it checks each character's Unicode metadata:

graph LR
    subgraph "Text: 'Hello Σ世界'"
        C1[H] --> C2[e]
        C2 --> C3[l]
        C3 --> C4[l]
        C4 --> C5[o]
        C5 --> C6[' ']
        C6 --> C7[Σ]
        C7 --> C8[世]
        C8 --> C9[界]
    end

    subgraph "Pattern: \\p{Greek}+"
        P["Check each
character"]
        P --> M1{Is Greek?}
    end

    C7 -."Unicode
property".-> M1
    M1 -->|Yes| R1[✓ Match Σ]

    style C7 fill:#e8f5e9
    style R1 fill:#c8e6c9

Figure: Unicode property matching examines character metadata. Only 'Σ' matches \p{Greek} in this example.

Common Unicode Properties¶

# Match any character including newlines without --multiline-dotall
rg '\p{any}+'

# Equivalent to using . with multiline-dotall
rg --multiline-dotall '.+'

Note on Emoji Matching: Emoji matching with \p{Emoji} may vary across different regex engines and Unicode versions. Some complex emoji (like multi-codepoint sequences, skin tone modifiers, or zero-width joiners) may require additional pattern logic. Always test emoji patterns with your specific use case and data.

For a comprehensive list of Unicode properties, see the Rust regex Unicode documentation.

graph LR
    subgraph Default["Default Engine (Rust regex)"]
        direction TB
        D1["Standard Unicode
        Properties"]
        D2["Fast Performance"]
        D3["UTF-8 Optimized"]
    end

    subgraph PCRE2["PCRE2 Engine"]
        direction TB
        P1["Extended Unicode
        Properties"]
        P2["Advanced Features"]
        P3["Perl Compatibility"]
    end

    Default -.->|"Use when"| U1["Standard properties
    sufficient"]
    PCRE2 -.->|"Use when"| U2["Need extended
    Unicode support"]

    style Default fill:#e1f5ff
    style PCRE2 fill:#fff3e0

Unicode-Aware Metacharacters¶

By default, these metacharacters are Unicode-aware:

• \w: Matches all Unicode word characters (not just ASCII)
• \s: Matches all Unicode whitespace
• \d: Matches Unicode decimal digits
• \b: Unicode word boundaries

# Match Unicode word characters (includes accented letters, etc.)
rg '\w+'

# Match Unicode whitespace
rg '\s+'

Case-Insensitive Unicode¶

The -i flag performs Unicode case folding:

# Matches "café", "CAFÉ", "Café", etc.
rg -i 'café'

# Works across all scripts
rg -i 'Σ'  # Matches Σ and σ (Greek)

Disabling Unicode¶

For ASCII-only searches with better performance, use --no-unicode:

# Matches all Unicode word characters
rg '\w+'  # Includes café, Σ, 世界, etc.

# ASCII-only mode (faster for ASCII text)
rg --no-unicode '\w+'  # Only [a-zA-Z0-9_]

Note: --no-unicode affects the entire search, not individual patterns.

Common Use Cases¶

Validating International Names¶

# Match names with Unicode letters (supports accents, non-Latin scripts)
rg '^\p{Alphabetic}+( \p{Alphabetic}+)*$'  # (1)!

# Find names containing specific scripts
rg '\p{Han}+\s+\p{Latin}+'  # Chinese + Latin names

1. Matches one or more alphabetic characters, followed by optional space-separated words (e.g., "José García", "李明", "Müller")

Processing Multilingual Text¶

# Extract sentences from mixed-script documents
rg '\p{Uppercase}\p{Alphabetic}+.*?[.!?]'  # (1)!

# Find all non-ASCII text
rg '[^\p{ASCII}]+'  # (2)!

# Identify specific language blocks
rg '\p{Arabic}+' --only-matching  # (3)!

1. Matches sentences starting with uppercase, containing alphabetic chars, ending with punctuation (works across all scripts)
2. Negated ASCII property matches any character outside the ASCII range (useful for finding internationalized content)
3. Extracts Arabic text blocks; use --only-matching to output just the matched text without file/line context

Data Validation¶

# Validate Unicode whitespace handling
rg '\p{White_Space}+' --replace ' '  # (1)!

# Find problematic characters
rg '[^\p{Print}\p{White_Space}]'  # (2)!

1. Normalize all types of Unicode whitespace (tabs, non-breaking spaces, etc.) to regular spaces
2. Find non-printable control characters that may cause display or parsing issues

Working with Emoji¶

# Find lines containing emoji
rg '\p{Emoji}'  # (1)!

# Extract emoji from text
rg '\p{Emoji}+' --only-matching  # (2)!

# Find text without emoji (using negative pattern in broader search)
rg '^[^\p{Emoji}]+$'  # (3)!

1. Matches any line containing at least one emoji character
2. Extracts sequences of emoji without surrounding context (useful for emoji inventories)
3. Matches entire lines that contain no emoji characters (useful for filtering emoji-free content)