How to Find and Replace Whole Words Without Changing Other Text

A quick replacement can change more than you intended. Searching for cat may alter category, and a case-insensitive search may replace both cat and Cat. Start with a small sample and choose the matching rules before editing an entire document.

The Find and Replace Text tool provides separate controls for case, whole words, regular expressions and replacing all matches. The examples below show how those choices affect the result.

Test a whole-word replacement

Paste this short sample into Your text:

cat category Cat concatenate

Set Find to cat and Replace with to dog. Enable Whole word only and leave Match case and Use regular expression off. Leave Replace all on, then press Replace.

Our tool check returned two replacements and this output:

dog category dog concatenate

The two longer words remain intact. The capitalized word becomes lowercase because the replacement is inserted exactly as entered; the tool does not preserve each match’s capitalization automatically.

Choose case and replacement count deliberately

Change only the lowercase occurrence

Enable Match case for the same sample. The exact lowercase word matches, while Cat remains unchanged. This is useful when capitalized terms have a different meaning or need separate editorial treatment.

Change only the first occurrence

Turn Replace all off to change just the first match in the input. Use this for a single correction or a cautious first test. Do not assume it means the first occurrence on every line: it means one match across the whole text block.

Remove a phrase

Leave Replace with empty to delete matches. Review the surrounding spaces afterward. Removing a phrase can leave double spaces or an incomplete sentence even when the matching itself was correct.

Keep literal searches separate from patterns

For an exact phrase, leave regular-expression mode off. A period, plus sign or parenthesis then belongs to your search text. In pattern mode, those characters can have special meanings. MDN’s regular expression guide explains how patterns and escaping work.

Whole-word mode uses JavaScript word boundaries. Treat it as a matching rule, not a language-aware word detector. Test punctuation, hyphens and non-English text separately rather than assuming every visible word has the same boundary behavior.

Reorder a simple two-part name with capture groups

For these deliberately simple sample records:

Aster, Mira
Vale, Noah

Enable Use regular expression, disable Whole word only, and enter this pattern directly in Find, without surrounding slashes:

(\w+), (\w+)

Set the replacement to $2 $1 and keep Replace all enabled. The two parenthesized matches supply the numbered groups. Our tool check returned:

Mira Aster
Noah Vale

This narrow pattern assumes one simple word on each side of a comma followed by one space. It is not a general name parser; multiword surnames, apostrophes and accented names need a different tested pattern. The tool supports numbered replacement groups in regex mode. MDN documents the underlying concept in its string replacement reference; do not assume every JavaScript replacement feature is implemented by this tool.

Review the result before copying it back

1. Keep an original copy of the text outside the tool.
2. Compare the replacement count with the number you expected.
3. Inspect a changed line, a line that should stay unchanged, and any punctuation around matches.
4. For a larger edit, compare original and output in the Text Compare tool.

If the input also contains repeated records, handle that as a separate cleanup step using our guide to removing duplicate lines while keeping their order. Keeping each operation separate makes unintended changes easier to spot.