How to grep or sed with two patterns and print result in one line-CodePudding

My bash variable $line contains output of other command. Ip and username could be in different places inside one line :

10.20.0.11 01User3645 123213123213 http://amazon.com
af 10.20.0.12 http://amazon.com 02User4536
01User3645 123213123213 10.20.0.11 http://amazon.com
af 02User4536 http://amazon.com 10.20.0.12
01User0011 123213123213 http://amazon.com 10.20.0.11

I'm trying to sed or grep this variable with two patterns to get only ip address and username:

10.20.0.11 01User3645
10.20.0.12 02User4536
10.20.0.11 01User3645
10.20.0.12 02User4536
10.20.0.11 01User0011

With commands bellow it was possible to print desired result but with separated lines:

echo "$line" | grep -E -o '10\.[0-9] \.[0-9] \.[0-9] |[0-9] User[0-9] '

Output:

10.20.0.11
01User3645
10.20.0.12
02User4536
10.20.0.11
01User3645

The question here is how to combine output of ip-address and username in one line for each line in variable? Thanks!

CodePudding user response：

With bash and two regex:

echo "$line" \
| while read -r line; do
    [[ $line =~ 10\.[0-9] \.[0-9] \.[0-9]  ]] && echo -n "${BASH_REMATCH[0]} ";
    [[ $line =~ [0-9]{2}User[0-9]{4}       ]] && echo    "${BASH_REMATCH[0]}";
  done

Output:

10.20.0.11 01User3645
10.20.0.12 02User4536
10.20.0.11 01User3645
10.20.0.12 02User4536
10.20.0.11 01User0011

CodePudding user response：

Here is an awk solution for this task:

awk '
match($0, /10\.[0-9] \.[0-9] \.[0-9] /) {
   printf "%s ", substr($0, RSTART, RLENGTH)
}
match($0, /[0-9] User[0-9] /) {
   print substr($0, RSTART, RLENGTH)
}' <<< "$line"

10.20.0.11 01User3645
10.20.0.12 02User4536
10.20.0.11 01User3645
10.20.0.12 02User4536
10.20.0.11 01User0011