Linux Bash/Sed中提取域名与端口的正则表达式需求

阿华AIGC实验室

2026-5-19

Extract Domain and Port in Bash/Sed Using Regex

Got it, let's work through this to get that clean domain:port combo you need. You already nailed matching the leading -epool (.*://)? section—now we just need to either strip that off or directly grab what comes right after it. Here are a few practical, Bash/Sed-friendly methods:

Method 1: Strip Leading Unwanted Content with Sed

If your input lines look like miner -epool http://lb.geo.pirlpool.eu:8002 -ewal mywallet, you can use Sed to erase the leading matched portion entirely, leaving only the domain:port.

Use this command (extended regex makes the syntax cleaner):

sed -E 's/.*-epool (.*:\/\/)?//' your_input_file.txt

Breakdown:

-E enables extended regular expressions (no need to escape parentheses)
.*-epool (.*:\/\/)? matches everything up to and including the -epool flag plus an optional http:///https:// prefix
Replacing that match with an empty string leaves exactly the domain:port you're targeting.

Method 2: Directly Capture the Target with Sed (Ideal for Lines with Extra Text)

If there's extra content after the domain:port, you can explicitly capture the non-space characters that form your target and replace the whole line with just that capture:

sed -E 's/.*-epool (.*:\/\/)?([^[:space:]]+).*/\2/' your_input_file.txt

Breakdown:

[^[:space:]]+ matches one or more non-space characters (this is your domain:port, since it won't contain spaces)
\2 refers to the second captured group (the domain:port), so we replace the entire line with just that value.

Method 3: Use Grep for Direct Extraction

If you prefer Grep, its -o flag lets you output only the matched portion. GNU Grep supports positive lookbehind, which lets you confirm the leading section exists without including it in the result:

grep -Eo '(?<=-epool (.*://)?)[^[:space:]]+' your_input_file.txt

Breakdown:

(?<=-epool (.*://)?) is a positive lookbehind—it verifies the leading -epool and optional prefix are present, but doesn't include them in the output
[^[:space:]]+ captures the domain:port, and -o makes Grep print only that part.