How to read list in JSON file using awk and sed

Question

I have a list in JSON file apps.json:

{
"properites":{
   "apps":["apple","orange","mango"]
}
}

I need to read this file and get of objects.

I used:

fwk_app = $( awk '/apps/{ print }' ${apps.json} | grep apps | awk -F: '{ print } substr($0, index($0,$2))}' |sed 's/(,*\)"[/\"]*$/\1/' );
for item in fwk_app:
do
echo"$item"
done

The output is:

mango

The output I need is:

apple
orange
mango

Answer 1

With awk you can get the expected output

awk -v FS=':' -v OFS='\n' '/"apps"/ {gsub(/[]"[]/,"",$2);split($2,a,/,/);print a[1],a[2],a[3]}' file
apple
orange
mango

but I think the correct way is @Kusalananda's solution.

Answer 2

The correct way to parse a JSON file is to use a JSON parser. The jq utility is one such parser that is made for use on the command line and in scripts:

jq -r '.properites.apps[]' apps.json

This would output

apple
orange
mango

This asks jq to extract the elements of the array apps in the top-level object properites in the JSON document in the file apps.json. The -r option asks jq to output "raw" data rather than JSON-encoded data.

To get tab-delimited output on a single line, use

jq -r '.properites.apps | @tsv' file

This would output

apple   orange  mango

The @tsv operator takes an array and outputs it as a tab-delimited list. You could instead use @csv to get properly quoted CSV output.

Note that this would also work no matter whether your JSON document was

{
  "properites": {
    "apps": [
      "apple",
      "orange",
      "mango"
    ]
  }
}

or

{"properites":{"apps":["apple","orange","mango"]}}

(these are equivalent forms of the same JSON document).

Answer 3

You could use python:

python -c '
import json, sys
data = json.load(sys.stdin)
for el in data["properites"]["apps"]:
  print (el)
' < apps.json

Answer 4

jq is the best answer. ruby ships with a JSON module:

ruby -rjson -e '
    data = JSON.parse(IO.read(ARGV.shift))
    puts data["properites"]["apps"].join("\n")
' apps.json

Answer 5

You could also use grep -P if available:

grep -wF  '"apps":' apps.json | grep -Po '[[,]"\K[^"]*'

However, this (and the awk solution provided by @Carlos) works only if the input is exactly as shown, not if the json with same content is formatted differently.

A proper json parser , e.g. jq or python's json module should be preferred!