Word and Line Boundaries¶

Boundary matching ensures you match complete words or lines, preventing partial matches that can clutter search results.

Whole Word Matching¶

Use -w or --word-regexp to match whole words only:

# Without -w
rg "test"           # Matches "test", "testing", "contest", "latest"

# With -w
rg -w "test"        # Matches only "test"

flowchart LR
    Text["Text: testing test contest latest"]

    NoFlag["rg 'test'
(no -w flag)"]
    WithFlag["rg -w 'test'
(with -w flag)"]

    NoFlag --> M1["✓ testing"]
    NoFlag --> M2["✓ test"]
    NoFlag --> M3["✓ contest"]
    NoFlag --> M4["✓ latest"]

    WithFlag --> M5["✗ testing"]
    WithFlag --> M6["✓ test"]
    WithFlag --> M7["✗ contest"]
    WithFlag --> M8["✗ latest"]

    Text --> NoFlag
    Text --> WithFlag

    style M1 fill:#ffebee
    style M2 fill:#e8f5e9
    style M3 fill:#ffebee
    style M4 fill:#ffebee
    style M5 fill:#f5f5f5
    style M6 fill:#e8f5e9
    style M7 fill:#f5f5f5
    style M8 fill:#f5f5f5
    style NoFlag fill:#fff3e0
    style WithFlag fill:#e1f5ff

Figure: Comparison of pattern matching with and without -w flag showing how word boundaries prevent partial matches.

Practical Use Cases¶

# Find the "log" function, not "logger" or "login"
rg -w "log"

# Find "test" in test names, not "latest" or "contest"
rg -w "test"

# Find "error" variable, not "errors" or "error_handler"
rg -w "error"

Why this matters: Without word boundaries, searching for common terms returns too many false positives, making it hard to find what you actually need.

Whole Line Matching¶

Use -x or --line-regexp to match entire lines only:

# Match lines that are exactly "TODO"
rg -x "TODO"

# Match lines containing only digits
rg -x "[0-9]+"

# Combined with other flags
rg -x -i "error"    # Match lines that are exactly "error" (any case)

This is equivalent to surrounding your pattern with ^ and $ line anchors.

# Find TODO comments that are on their own line
rg -x "TODO"

# Find configuration lines with just numbers (like port numbers)
rg -x "[0-9]+"

# Find lines that are exactly a specific keyword
rg -x -i "error"    # Case-insensitive exact match

Why this matters: Line boundaries help you find complete configuration entries, standalone markers, or lines with specific formats, filtering out lines where the pattern appears as part of larger content.

flowchart TD
    Start[rg command with flags]

    HasBoth{Has both -x and -w?}
    HasX{Has -x flag?}
    HasW{Has -w flag?}

    Start --> HasBoth
    HasBoth -->|Yes| CheckOrder{Which comes last?}
    HasBoth -->|No| HasX

    CheckOrder -->|-x last| UseX[Use -x: Line matching]
    CheckOrder -->|-w last| UseW[Use -w: Word matching]

    HasX -->|Yes| UseX
    HasX -->|No| HasW

    HasW -->|Yes| UseW
    HasW -->|No| Normal[Normal pattern matching]

    UseX --> Result1["Matches entire lines only"]
    UseW --> Result2["Matches whole words only"]
    Normal --> Result3["Matches pattern anywhere"]

    style UseX fill:#ffebee
    style UseW fill:#e1f5ff
    style Normal fill:#f5f5f5
    style Result1 fill:#ffebee
    style Result2 fill:#e1f5ff
    style Result3 fill:#f5f5f5

Figure: Flag precedence decision flow showing how -x and -w interact. When both are specified, the last flag wins.

rg -w -x "test"   # Uses -x (line matching) - last flag wins
rg -x -w "test"   # Uses -w (word matching) - last flag wins

Word Boundaries in Regex¶

Basic Boundary Matching¶

# \b matches word boundary
rg "\btest\b"        # Matches "test" but not "testing" or "contest"

# \B matches non-word boundary
rg "\Btest"          # Matches "contest" but not "test"
rg "test\B"          # Matches "testing" but not "test"

Unicode Word Boundaries¶

Word boundaries work correctly with Unicode characters, treating them according to Unicode word boundary rules:

# Works with Unicode words
rg -w "café"         # Matches "café" as a complete word

# Unicode-aware boundaries in regex
rg "\bcafé\b"        # Matches complete Unicode word "café"

# Handles grapheme clusters correctly
rg -w "naïve"        # Matches "naïve" but not "naive"

Advanced Boundary Examples¶

# Match word starting with prefix
rg "\btest\w*"       # (1)!

# Match word ending with suffix
rg "\w*ing\b"        # (2)!

# Negative word boundary - match within words only
rg "\Btest\B"        # (3)!

1. \b ensures start of word, \w* matches remaining word characters. Matches "test", "testing", "tests"
2. \w* matches any word characters before "ing", \b ensures end of word. Matches "testing", "running", "coding"
3. \B matches non-word boundaries on both sides. Only matches "test" when surrounded by word characters (e.g., "contest")