CODEWITHSUNDEEP
jsonjq
DarkSide
DarkSide

Reputation: 3709

Replace some keys in json by jq

I have json file like this one:

{
    "abc": "A",
    "abc-release": "B",
    "bcd":"C",
    "cde-release":"D",
    "cde":"E"
}

and would like to get resulting file like this:

{
    "abc": "B",
    "bcd":"C",
    "cde":"D"
}

That is:

  1. remove all elements with key X if we also have key X-release;
  2. rename elements with key X-release as X;

Need to make this work by using Linux commands like jq and sed. Preferably as one-liner with pipes.

I have spent multiple hours trying to implement this and best I could do is this jq command which returns array of keys which i need to delete.

jq "with_entries(select(.key|endswith(\"-release\"))) | keys | map(split(\"-\")[0])"

[
    "abc",
    "cde"
]

But how to feed this array in jq del method or some other way to get result what I need?

Second step (renaming elements) is easy. That can be done by jq but also can be done at the end simply by sed:

// simply like this
sed 's/-release//'
// or
sed 's/(.*)-release/$1/'

Upvotes: 1

Views: 172

Answers (1)

peak
peak

Reputation: 116670

The following constructive solution using jq has the advantage of being simple and efficient:

. as $in
| reduce (keys_unsorted[] | select(endswith("-release")|not)) as $k ({};
    . + {($k) : (($k + "-release") as $kr
                 | $in
                 | if has($kr) then .[$kr] else .[$k] end) } )

Upvotes: 3

Related Questions