Grep value from json file

Question

my json file looks something like this

[{"product":"Apple","id":"2134"},{"product":"Mango","id":"4567"}]

I want to grep/sed/awk 'id' for given 'product'.

Output:

Enter product : Apple
Your product id is : 2134

Answer 1

Use a JSON aware tool. Perl has the JSON library:

#!/usr/bin/perl
use warnings;
use strict;

use JSON;

my $json = '[{"product":"Apple","id":"2134"},{"product":"Mango","id":"4567"}]';

print 'Enter product: ';
chomp( my $product = <> );

print 'Your product id is: ', 
    (grep $_->{product} eq 'Apple', @{ from_json($json) })[0]{id}, "\n";

Answer 2

Use a json parser, not sed/grep/awk.

Using Python json module:

#!/usr/bin/env python2
import json
with open('file.json') as f:
    f_json = json.load(f)
    print 'Enter product : ' + f_json[0]['product'] + '\nYour product id is : ' + f_json[0]['id']

Output:

Enter product : Apple
Your product id is : 2134

Answer 3

I created a file called json that looks like this:

[{"product":"Apple","id":"2134"},{"product":"Mango","id":"4567"},{"product":"Pear","id":"1111"},{"product":"Banana","id":"2222"}]

Then I run this on the command line:

 cat json | sed -e 's/.\?{"product"\:\"\([^\"]\+\)\","id"\:\"\([[:digit:]]\+\)[^}]}\+.\?/Enter Product : \1\nYour Product id is : \2\n/mgi'

The output is this:

Enter Product : Apple
Your Product id is : 2134
Enter Product : Mango
Your Product id is : 4567
Enter Product : Pear
Your Product id is : 1111
Enter Product : Banana
Your Product id is : 2222

Answer 4

sed is clumsy, but here is what I could get -

sed  's/.*\(Apple\)[^:]*:"\([0-9]*\)".*/Enter product ID: \1\nYour  product ID is: \2/' 3
Enter product ID: Apple
Your product ID is: 2134

sed  's/.*\(Mango\)[^:]*:"\([0-9]*\)".*/Enter product ID: \1\nYour product ID is: \2/' 3
Enter product ID: Mango
Your product ID is: 4567

Edited based on latest input which is slightly different

./prod.sed Apple
Enter product ID: Apple
Your product ID is: 1234

./prod.sed Mango
Enter product ID: Mango
Your product ID is: 12345

prod.sed (use 3rd line for publishName)

#sed  's/.*product_id":"\(Apple\).*"productBuildId":"\([0-9]*\)".*/Enter product ID: \1\nYour product ID is: \2/' data
sed  's/.*product_id":"\('"$1"'\).*"productBuildId":"\([0-9]*\)".*/Enter product ID: \1\nYour product ID is: \2/' data
sed  's/.*product_id":"\('"$1"'\)","publishName":"\([^"]*\)".*/Enter product ID: \1\nYour publish ID is: \2/' data

explanation : substitute the string <anything> Mango <anything other than :> : <number> with -> Your product ID is: Mango <next line> 
Your product ID is : <the number we got> 

 \1 and \2 save the matched expression in \(..\) for later use.

Answer 5

Given a product name in $name, the ID of that product could be parsed out using jq like so:

jq -r --arg name "$name" '.[] | select(.product == $name).id' file.json

Example:

$ cat file.json
[{"product":"Apple","id":"2134"},{"product":"Mango","id":"4567"}]

$ name=Apple
$ jq -r --arg name "$name" '.[] | select(.product == $name).id' file.json
2134

Answer 6

Given your input, the following sed command ...

sed -e 's/[}"]*\(.\)[{"]*/\1/g;y/,/\n/' <your_input

...will write to output...

[product:Apple
id:2134
product:Mango
id:4567]

...probably you can take it from there.