如何按字典序对提取出的IP地址进行排序以匹配预期输出

The issue here is that the default sort command uses lexicographical (string) sorting, which compares characters one by one rather than treating each IP segment as a numeric value. That's why you see results like 127.0.0.1 appearing before 7.7.7.8—the string "1" has a lower ASCII value than "7", even though 7 is numerically smaller than 127.

Simple Solution: Use GNU Sort's Version Sort (sort -V) #

If you're on a system with GNU sort (most Linux distributions), the easiest fix is to use the -V (version sort) flag. This option is designed specifically for sorting strings with numerical segments (like IP addresses, version numbers) by treating each segment as a number:

find /root/data -type f -exec grep -E -o "(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)" {} \; | sort -V | uniq -c | awk '{print $2}'

sort -V will automatically parse each IP segment as a numeric value, sorting them in the order you expect: starting with 0.0.0.0, then 7.7.7.8, followed by the 15.x, 26.x, etc., ranges, and finally the 127.x and higher segments.

Alternative for Non-GNU Systems (e.g., BSD/macOS) #

If sort -V isn't available, you can use awk to normalize each IP segment to a fixed 3-digit width (padding with leading zeros), which makes lexicographical sorting match numerical sorting. Here's how:

find /root/data -type f -exec grep -E -o "(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)" {} \; | uniq | awk -F. '{printf("%03d.%03d.%03d.%03d\t%s\n", $1,$2,$3,$4,$0)}' | sort | cut -f2

How this works: #

• uniq first removes duplicates (we moved this earlier to avoid processing redundant entries)
• awk formats each IP with leading zeros (e.g., 7.7.7.8 becomes 007.007.007.008), paired with the original IP
• sort now works correctly on the padded string
• cut extracts the original (unpadded) IP from the sorted output

This will produce the exact ordered list you're expecting.

内容的提问来源于stack exchange，提问作者Patrick