How do I search or substitute only within specific columns of text in Vim?

Explanation

The \%Nc, \%>Nc, and \%<Nc atoms let you anchor a search pattern to a specific column position. This makes it possible to match text only when it appears at, after, or before a given column — a powerful but rarely-used feature of Vim's regex engine.

How it works

• \%20c — match only at column 20
• \%>20c — match only after column 20 (i.e., from column 21 onward)
• \%<20c — match only before column 20 (i.e., up to column 19)

Column numbers are 1-based and count byte positions by default. Combine the atom with a pattern to restrict where a match can occur:

/\%>40cfoo

This finds foo only when it starts past column 40 — ignoring any occurrence of foo in the first 40 columns.

Example

Suppose you have a log file where timestamps occupy columns 1–24 and the message starts at column 25. To highlight the word ERROR only in the message field (not accidentally in a timestamp):

2024-01-15 10:42:00 UTC ERROR something failed

/\%>24cERROR

Or to substitute only within columns 25–80:

:%s/\%>24c\%<81cERROR/WARN/g

Tips

• Combine \%>Nc and \%<Mc in a single pattern to create a column range.
• The byte-column \%c and character-column \%v variants exist: \%v respects multi-byte characters and 'virtualedit', making it the right choice for unicode-heavy buffers.
• This works inside :substitute, :global, and any other command that accepts a search pattern.