Iterate over some pattern using shell

Question

I have a file to process and get some entries from it. The file format is:

Example Input:

String1:"Hello"
String2:"World"
String3:"Question"
End
String1:"HelloAgain"
String5:"Answer"
End
String1:"NewEntry"
String2:"Foo"
String5:"NewFile"
End

Script will iterate and try to find String1 and print it, if String2 also present in that block append it to String1 and print.

Example Output:

Hello World
HelloAgain
NewEntry Foo

I need to loop through every such block.What I mean by a block here is content between every "End" string. I have a script like:

    cat $1 | while read line
do
    if [[ $line == "String1="* ]];
    then
        string1=$line
    fi
    if [[ $line == "String2="* ]];
    then
        string2=$line
    fi
    if [[ $line == "End" ]];
    then
        if [ $string1 ];
            then
            echo "string1/"$string1"
        fi
        if [ $string1 ] && [ $string2 ];
            then
            echo $string1" "$string2
        fi
    #Reset values
    string1=''
    string2=''
    fi
done

This code works fine but the processing in between these blocks can become more complex, whats the best way to iterate such blocks and process those block entries.

NOTE:The file is kind of a property file.

Answer 1

Given your provided input and requested output:

$ awk -F: '$0 == "End" { printf "\n" } $1 ~ /^String[12]$/ { gsub( /\"/, "", $2 ); printf( "%s ", $2)}' input
Hello World
HelloAgain
NewEntry Foo

Answer 2

perl -F':|"' -lane 'if ($out && $_ eq "End") {print "$out"; $out=""};
  $out  = $F[2] if ($F[0] eq "String1");
  $out .= $F[2] if ($out && $F[0] eq "String2")' input.txt

This perl script uses either : or " as the field separator, to avoid the need to strip double-quotes from the output. this means we need to extract the third field rather than the second as there is now an empty field between the first : and the ". Perl arrays are zero-based, so that's $F[2].

It only prints the output for a record ("block") if there is something to print, either String1 by itself or String1 and String2 are present - i.e. it doesn't print blank lines for non-matching records.

Here's the same algorithm in awk:

awk -F':|"' 'out && $0 == "End"  { print out; out="" };
  $1 == "String1" { out=$3 };
  out && $1 == "String2" { out = out$3 }' input.txt