How to convert linefeed into literal "\n" - json

Question

I'm having some trouble converting my file to a properly formatted json string.
Have been fiddling with sed for ages now, but it seems to mock me.
Am working on RHEL 6, if that matters.
I'm trying to convert this file (content):
Hi there...
foo=bar
tomàto=tomáto
url=http://www.stackoverflow.com
Into this json string:
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com"}
How would I replace the actual line feeds in the literal '\n' character?? This is where I'm utterly stuck!
I've been trying to convert line feeds into ";" first and then back to a literal "\n". Tried loops for each row in the file. Can't make it work...
Some help is much appreciated!
Thanks!

Answer 1

sed is for simple substitutions on individual lines, that is all. Since sed works line by line your sed script doesn't see the line endings and so you can't get it change the line endings without jumping through hoops using arcane language constructs and convoluted logic that hasn't been useful since the mid-1970s when awk was invented.
This will change all newlines in your input file to the string \n:
$ awk -v ORS='\\n' '1' file
Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com\n
and this will do the rest:
$ awk -v ORS='\\n' 'BEGIN{printf "{\"text\":\""} 1; END{printf "\"}\n"}' file
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com\n"}
or this if you have a newline at the end of your input file but don't want it to become a \n string in the output:
$ awk -v ORS='\\n' '{rec = (NR>1 ? rec ORS : "") $0} END{printf "{\"text\":\"%s\"}\n", rec}' file
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com"}

Answer 2

With GNU sed:
sed ':a;N;s/\n/\\n/;ta' file | sed 's/.*/{"text":"&"}/'
Output:
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com"}

Answer 3

Use awk for this :
awk -v RS=^$ '{gsub(/\n/,"\\n");sub(/^/,"{\"text\":\"");sub(/\\n$/,"\"}")}1' file
Output
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com"}

Answer 4

awk to the rescue!
$ awk -vRS='\0' '{gsub("\n","\\n");
print "{\"text\":\"" $0 "\"}"}' file
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com\n"}

Answer 5

This might work for you (GNU sed):
sed '1h;1!H;$!d;x;s/.*/"text":"&"/;s/\n/\\n/g' file
Slurp the file into memory and use pattern matching to manipulate the file to the desired format.

Answer 6

The most simple (and elegant ?) solution :) :
#!/bin/bash
in=$(perl -pe 's/\n/\\n/' $1)
cat<<EOF
{"text":"$in"}
EOF
Usage:
./script.sh file.txt
Output :
{"text":"Hi there...\n\nfoo=bar\ntomàto=tomáto\nurl=http://www.stackoverflow.com\n"}