jsondepth

A small command-line tool to walk through the depth levels of a json objects

output json | jd <PATH>=. <DEPTH>=0

Summary

• Motivation
• Installation
• How-to
  • Specify a depth level
  • Specify a path
  • Specify a depth and a path
  • Access JS objects attributes
    • Native attributes
      • Arrays length
    • Special keys
      • _keys
      • _values
      • _map
      • _first
      • _last
  • Format the output as valid JSON
  • Parse newline delimited JSON
• See also

Motivation

Working with super deep json objects from the terminal is a pain, unless you use a good json parser.

jq is an awesome one, but doesn’t handle object depths, afaik.

Here the idea is to walk through a json object as you would read a summary: level by level.

Installation

npm install -g jsondepth

How-to

real world example:

url='https://www.wikidata.org/w/api.php?action=wbgetentities&ids=Q1&format=json'
curl -s "$url" | jsondepth

logs the object with only the first-level keys and values:

{ entities: [Object], success: 1 }

for the sake of convenience and lazyness, jsondepth is aliased to jd


(which, purposedly make it look a bit like jq)

curl -s "$url" | jd

Specify a depth level

this is equivalent to the two previous one:

curl -s "$url" | jd 0

now let’s go one-level deeper:

curl -s "$url" | jd 1

outputs:

{ entities: { Q1: [Object] }, success: 1 }

curl -s "$url" | jd 2
curl -s "$url" | jd 3
curl -s "$url" | jd 4
# etc

curl -s "$url" | jd -1

Specify a path

curl -s "$url" | jd entities.Q1.aliases.fi.1
# or to mimick jq syntax
curl -s "$url" | jd .entities.Q1.aliases.fi.1
# or with keys with spaces
curl -s "$url" | jd .entities.Q1.['a key with spaces']

Specify a depth and a path

if both a path and a depth are specified, path should come first, depth second

curl -s "$url" | jd entities.Q1.claims.P31.0 3

Access JS objects attributes

Native attributes

Arrays length

curl -s "$url" | jd entities.Q1.aliases.en.length
# => 5

Special keys

_keys

apply Object.keys to the final object

curl -s "$url" | jd entities._keys
# => ['Q1']

_values

apply _.values to an array

curl -s "$url" | jd entities._values
# => [{ id: 'Q1', etc...}]
curl -s "$url" | jd entities._values.0.id
# => 'Q1'

_map

map properties of an array

curl -s "$url" | jd entities._values._map.id
# => ['Q1']

_first

echo '["foo", "bar"]' | jd ._first
# => foo

_last

echo '["foo", "bar"]' | jd ._first
# => bar

Format the output as valid JSON

Wrapped results like { entities: { Q1: [Object] }, success: 1 } are more readible but aren’t valid JSON. And sometimes, for whatever reason, you want a valid JSON output; there in a option for that: --json|-j

curl -s "$url" | jd entities.Q1.aliases --json

You can even specify the output indentation level (Between 0 and 9. Default: 2):

curl -s "$url" | jd entities.Q1.aliases --json=0
curl -s "$url" | jd entities.Q1.aliases --json=4

Notice that it disables the depth option as it’s responsible for the wrapping mechanism.

Parse newline delimited JSON

Add the --line option to parse the stdin line by line

curl https://some.ndjson.dump | jd --line .key

There is a tolerance for JSON lines ending by a comma, and any line that isn’t starting by a { or a } is ignored.

See also

• jq
• fx