使用 sed 和正则表达式从文件中删除文本

Question

I'm trying to use sed with regex to remove specific patterns from a file. Here is the command I'm using:

sed '/location .*?\/\s+{(?:[\w\W]+?)}\s*(?=(?:location|$))/d'

Here is a sample text I'm testing with:

location / {
                try_files $uri $uri/ /index.php?$args;
}
location ~*  .(jpg|jpeg|png|gif|ico|css|js)$ {
            expires 365d;
        }

        location ~*  .(pdf)$ {
            expires 30d;
        }

        location ~ \.php$
                {
                        try_files $uri =404;
                        fastcgi_split_path_info ^(.+\.php)(/.+)$;
                        include /etc/nginx/fastcgi_params;
                        fastcgi_index index.php;
                        fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
                        include /etc/nginx/nginx_limits.conf;
                        if (-f $request_filename)
                        {
                                fastcgi_pass unix:/usr/local/php73/sockets/rgeoipdsm1.sock;
                        }
                }
        location ~ / {
                        try_files $uri $uri/ /index.php?$args;
        }

        location ^ / {
                        try_files $uri $uri/ /index.php?$args;
        }

        location ~ / {
                        if (-f $request_filename)
                        {
                                fastcgi_pass unix:/usr/local/php73/sockets/rgeoipdsm1.sock;
                        }
                        try_files $uri $uri/ /index.php?$args;
        }

location ~ \.png {

}

location ~ / {
}

What I'm expecting to see after running sed:

location ~*  .(jpg|jpeg|png|gif|ico|css|js)$ {
        expires 365d;
    }
    
    location ~*  .(pdf)$ {
        expires 30d;
    }

        location ~ \.php$
                {
                        try_files $uri =404;
                        fastcgi_split_path_info ^(.+\.php)(/.+)$;
                        include /etc/nginx/fastcgi_params;
                        fastcgi_index index.php;
                        fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
                        include /etc/nginx/nginx_limits.conf;
                        if (-f $request_filename)
                        {
                                fastcgi_pass unix:/usr/local/php73/sockets/rgeoipdsm1.sock;
                        }
                }
    location ~ \.png {

}

But after running sed I get no changes at all, it seems to match nothing. Is it possible to fine-tune this command (perhaps with escape characters?) so it'll remove the required patterns?

Answer 1

This might work for you (GNU sed):

sed -E '/location ([~^] )?\/ \{/{:a;N;/^(\s*)location.*\n\1\}\s*$/d;ba}' file

Match the location required, then append lines until the white space before the closing } matches the opening lines initial white space and delete those line.

Answer 2

You can use

sed '/^[[:blank:]]*location .*\/[[:blank:]]*{[[:blank:]]*$/,/^[[:blank:]]*}[[:blank:]]*$/d' yourfile > newfile

This sed command finds blocks of lines between line matching ^[[:blank:]]*location.*\/[[:blank:]]*{[[:blank:]]*$ and a line matching ^[[:blank:]]*}[[:blank:]]*$ , and removes them (with d ).

The ^[[:blank:]]*location.*\/[[:blank:]]*{[[:blank:]]*$ regex matches

• ^ - start of string
• [[:blank:]]* - zero or more spaces/tabs
• location - a location string
• .* - any zero or more chars
• \/ - a / char
• [[:blank:]]*{[[:blank:]]* - a `{~ enclosed with zero or more spaces/tabs
• $ - end of string.

The ^[[:blank:]]*}[[:blank:]]*$ regex matches a string that only contains a single } char eclosed with optional spaces/tabs.