The following two sed commands works.
sed -E "s#Example(.*)#\nconst ex='Example\1';\n#; s#Input:(.*)#const\1;#" in.txt | sed '/\S/!d'
However, how to rewrite them so that only one sed command is used? i.e.
sed -E "???" in.txt
I tried this but it did not work:
sed -E "s#Example(.*)#\nconst ex='Example\1';\n#; s#Input:(.*)#const\1;#; /\S/!d" in.txt
in.txt:
Example 1:
Input: s = "()"
GNU sed version is 4.7.
Expected output:
const ex='Example 1:';
const s = "()";
You are inserting line breaks into the pattern buffer, and the pattern buffer then has a multiline string in it. The /\S/!d command will not match a pattern buffer that contains a nonempty line with empty lines, so it will not delete empty lines from your multiline strings.
Fix is simple:
$ sed -E 's#Example(.*)#const ex='Example\1';#; s#Input:(.*)#const\1;#; /\S/!d' in.txt
const ex=Example1;
const s = "()";
... just don't add \n... sed is going to preserve those line breaks from the original line when you do s### substitution. You don't need to add more.
Also, in my shell ! was expanded inside double quotes, so I switched to single quotes.
I dunno with sed (line oriented), but with awk:
awk '
$1=="Example"{print "const ex=\047Example "$2"\047"}
$1=="Input:"{sub(/Input:/, "const ");print}
' file
Output:
const ex='Example 1:'
const s = "()"
AnotherExample Input "10"?