如何在Bash中用AWK匹配并打印所有符合规则的句子

Got it, let's fix this! The issue with your current awk command is that match() only finds the first matching sentence in the input string—without a loop to keep searching the remaining text, you'll only ever get that first result. Here's how to modify it to capture all valid sentences that meet your criteria:

Version 1: Output all valid sentences in a single line (matches your expected output) #

text="incorrect sentence! this is not sentence! This is sentence. this is not sencence. This is correct sentence."
echo "$text" | awk '{
    output = ""
    while (match($0, /[A-Z][^!?.]*[!?.]/)) {
        # Extract the matched valid sentence
        sentence = substr($0, RSTART, RLENGTH)
        # Append to output with space separation
        output = output (output ? " " : "") sentence
        # Trim the input string to start after the current match
        $0 = substr($0, RSTART + RLENGTH)
    }
    print output
}'

Output for your second example:

This is sentence. This is correct sentence.

Version 2: Output each valid sentence on its own line #

If you prefer individual lines instead of a single concatenated string, use this simpler loop:

echo "$text" | awk '{
    while (match($0, /[A-Z][^!?.]*[!?.]/)) {
        print substr($0, RSTART, RLENGTH)
        $0 = substr($0, RSTART + RLENGTH)
    }
}'

Output for your first example:

This is sentence!
This is sentence.

How it works: #

1. The while (match(...)) loop runs repeatedly as long as a valid sentence is found in the remaining text.
2. substr($0, RSTART, RLENGTH) uses the positions stored by match() to extract the exact matched sentence.
3. $0 = substr($0, RSTART + RLENGTH) updates the input string to skip the part we've already processed, so the next iteration looks for the next valid sentence.
4. The regex /[A-Z][^!?.]*[!?.]/ specifically targets strings that:
   • Start with an uppercase letter
   • Follow with any characters except !, ?, or .
   • End with one of those three punctuation marks

内容的提问来源于stack exchange，提问作者Amateur