JQ: Replace keys with corresponding value in another file

Question

I'm trying to replace the value of a key in my original JSON with the value of an object with the corresponding key in another document.

These are my two files: File 1:

{
    "-KaM0otlgWxXniYiacFe": {
        "-LNxx1IiX6oYTxJ4IXx2": true
    },
    "-KlJTvbfonIMI_-YfS5R": {
        "-LNxx1IbaB-yrZ623hrX": true
    }
}

File 2:

{
    "-KaM0otlgWxXniYiacFe": {
        "a": "-L-b__nH9PlMcz0stDDE",
        "b": "-L7ZNKSZy570TlrQUzHM",
        "c": "-Kaae3MsQUyViCKPs8Iv"
    },
    "-KlJTvbfonIMI_-YfS5R": {
        "a": "-LAlXKfUUTdYDeCZH-u-",
        "b": "-L7ZNKSTnob7w0HXjHr6",
    "    c": "-KYYicPD7VA9DEF_rus3"
    }
}

The goal is to create 3 new files, where the original keys have been replaced with the value of a, b and c in each file.

Desired result when looking against "a":

{
    "-L-b__nH9PlMcz0stDDE": {
        "-LNxx1IiX6oYTxJ4IXx2": true
    },
    "-LAlXKfUUTdYDeCZH-u-": {
        "-LNxx1IbaB-yrZ623hrX": true
    }
}

I have tried using something along the lines of:

cat file1.json | jq --slurpfile file2 file2.json '| map(with_entries(.key = .file2[.key].a'

But I'm feeling rather clueless as I havn't used jq before.

Any help is appreciated.

Update:

How do I handle the scenario where the key from File 1 might not exist in File 2, resulting in the "Cannot use null (null) as object key"?

oguz ismail · Accepted Answer · 2018-10-25T14:39:48.447

2

Not an expert on jq, but here is what I've come up with:

jq --arg k a --argfile file2 file2.json 'with_entries(.key |= $file2[.][$k])' file1.json > out_a.json

Change --arg k a and out_a.json for other keys and output files.

edited Oct 25 '18 at 14:39

answered Oct 25 '18 at 14:24

oguz ismail

1
16
47
69

1

The jq filter can be simplified: 'with_entries(.key |= $file2[.][$k])’ – peak Oct 25 '18 at 14:32
Thank you for the solution. How may I handle cases where the key may be non existent, null in File2? – buddhabath Oct 25 '18 at 15:08
Hmm, I have no idea. – oguz ismail Oct 25 '18 at 15:19

peak · Answer 2 · 2018-10-26T07:58:32.463

If you have more than a trivial number of keys "a", "b", ..., and if you don't mind if the output files are minified, you could avoid the multiple calls to jq using the approach illustrated here:

< file1.json jq -rc --argjson keys '["a", "b", "c"]' --argfile file2 <(file2) '
  $keys[] as $k 
  | ($k, (with_entries(.key |= $file2[.][$k]) ))
' | awk 'file {print > file ".json"; file=0; next} {file=$0}'

You could go one step further and compute $keys programatically using (for example):

   [$file2[]][0]|keys_unsorted

peak · Answer 3 · 2018-10-26T07:26:03.617

2

How do I handle the scenario where the key from File 1 might not exist in File 2, resulting in the "Cannot use null (null) as object key"?

The most obvious thing to do would probably be to test whether the object has the key:

with_entries(.key |= (if $file2[.]|has($k) then $file2[.][$k]  else . end))

edited Oct 26 '18 at 07:26

answered Oct 25 '18 at 16:39

peak

105,803
17
152
177

JQ: Replace keys with corresponding value in another file

3 Answers3

Linked