Skip to content

lesson_file_line_number_corruption.md

Latest commit

 

History

History
81 lines (54 loc) · 2.72 KB

lesson_file_line_number_corruption.md

File metadata and controls

81 lines (54 loc) · 2.72 KB
Raw
title File Content Corrupted by Terminal Line-Number Prefixes
domain devops
tags
terminal
sed
line-number
file-corruption
html
debug
status published
created 2026-06-20
source codewhale

Problem

You saved a file's content from terminal output (using cat -n, sed -n '...p', or less -N), but the file won't render/compile/run correctly. HTML pages show blank white. Scripts fail with syntax errors. JSON parsers reject the file.

Root Cause

Many terminal pager/display commands prefix each line with a line number when showing file content:

$ cat -n index.html
     1	<!DOCTYPE html>
     2	<html>
     3	<head>

If you copy-paste or redirect this output to a new file, every line now starts with 1|, 2|, 3| — making the file structurally invalid.

In the HTML case, the browser sees 1|<!DOCTYPE html> instead of <!DOCTYPE html>, which is not recognized as a valid doctype. The entire page falls into quirks mode and renders nothing.

For JavaScript/Python files, the line-number prefix causes syntax errors at the very first line.

Solution

Fix an already-corrupted file

Use sed to strip leading spaces + digits + pipe:

sed -i 's/^ *[0-9]*|//' corrupted-file.ext

This removes patterns like 1|, 23|, 100| from every line.

Verify before/after

# Before — inspect raw first line
head -1 corrupted-file.ext | od -c | head -3

# After — confirm first line is clean
head -1 corrupted-file.ext

Prevent it going forward

Instead of cat -n or sed -n '...p', use one of these clean alternatives:

Goal Safe command Why
View with line numbers cat -n file (visual only) Never redirect this to a file
Extract specific lines awk 'NR>=10 && NR<=20' file No line-number prefixes added
Copy a file cp file new-location Obvious but easy to forget
Save command output Redirect with > or tee Avoid using -n flags

If you must use sed with numbered lines

If you already captured a file with line prefixes and need to clean it in-place:

sed -i 's/^[[:space:]]*[0-9]*|//' file

This is the most portable version (works on BSD/macOS sed too).

Verification

# Check that no lines start with digits+pipe
grep -cP '^\s*\d+\|' file
# Should output 0

Notes

  • This also applies to YAML files — line-number prefixes break indentation-based parsing entirely.
  • In CI/CD logs captured from cat -n output, the same issue can cause downstream artifact corruption.
  • Most dangerous scenario: saving any file from a terminal session that used less -N, cat -n, sed -n '...p', or vi -c 'set number'.