How do I delete all lines between two delimiter patterns without removing the delimiters themselves?

Explanation

By combining the :global command with a relative address range, you can delete the content between repeating delimiter pairs while leaving the delimiters intact. The +1 offset skips the opening delimiter, and the -1 offset stops before the closing one.

How it works

• :g/BEGIN/ — for each line matching BEGIN
• +1 — start the range one line below the match (skip the opener)
• ,/END/ — search forward from the match for the next line matching END
• -1 — stop one line before that closing delimiter
• delete — delete the resulting range of lines

Example

Given:

BEGIN
  line to remove
  another line
END
BEGIN
  more content
END

Running :g/BEGIN/+1,/END/-1 delete results in:

BEGIN
END
BEGIN
END

Both content blocks are removed while BEGIN and END lines are preserved.

Tips

• Swap delete for s/.*// to blank lines instead of removing them
• Use d _ instead of delete to avoid polluting the unnamed register
• Replace BEGIN/END with any anchored pattern, e.g. ^### for Markdown section headers
• Works for multiple non-nested pairs in a single pass
• If BEGIN and END are adjacent, the range collapses and no lines are deleted (safe no-op)