使用bash从文本中提取正则表达式组并输出到文件

Question

I need to scan a log file and extract the relevant parts from it to another file. The log format is:

   [hh:mm:ss] Header
   [hh:mm:ss] irrelevant text
   [hh:mm:ss] irrelevant text
   [hh:mm:ss]Error text
   [hh:mm:ss] some details
   [hh:mm:ss] end_error;
   [hh:mm:ss] irrelevant text
   [hh:mm:ss] Warning text
   [hh:mm:ss] some details
   [hh:mm:ss] end_warning;
   [hh:mm:ss] irrelevant text
   [hh:mm:ss] irrelevant text
   [hh:mm:ss]Error text
   [hh:mm:ss] some details
   [hh:mm:ss] end_error;

I need to get all occurrences of Error and Warning and capture the following text:

[hh:mm:ss]Error text
[hh:mm:ss] some details
[hh:mm:ss] end_error;
[hh:mm:ss] Warning text
[hh:mm:ss] some details
[hh:mm:ss] end_warning;
[hh:mm:ss]Error text
[hh:mm:ss] some details
[hh:mm:ss] end_error;

What is the simplest way to achieve this on bash?

Answer 1

$ awk '/^(Error|Warning)/{f=1} f; /;/{f=0}' file
Error text
end_error;
Warning text
end_warning;

Your original input file showed Error and Warning at the start of each line so my script above has a start-of-line anchor (^) in it. Using your latest posted sample input file and desired output you'd need:

$ awk '
   /^[[:space:]]*\[[^]]+\][[:space:]]*(Error|Warning)/ { found=1 }
   found { sub(/^[[:space:]]+/,""); print }
   /;/ { found=0 }
' file
[hh:mm:ss]Error text
[hh:mm:ss] some details
[hh:mm:ss] end_error;
[hh:mm:ss] Warning text
[hh:mm:ss] some details
[hh:mm:ss] end_warning;
[hh:mm:ss]Error text
[hh:mm:ss] some details
[hh:mm:ss] end_error;

The complexity of the regexp is to avoid false matches if the words Error or Warning appear elsewhere in your input file.

Answer 2

Using GNU sed range operator with -n and -r option to suppress default printing and enabling extended regular expression respectively. p flag prints the line that matches the condition.

$ sed -nr '/^(Error|Warning)/,/;/p' file
Error text
end_error;
Warning text
end_warning;

You can do the same in awk too. But using Ed's approach is almost always recommended.

$ awk '/^(Error|Warning)/,/;/' file
Error text
end_error;
Warning text
end_warning;

Answer 3

Try:

cat file | awk '/^(Error|Warning)/,/;$/ { print $0 }' > output

This will pipe the file through awk, awk will print lines starting with Error or Warning up to the first line ending with ; , the result will be saved on output