使用JQ创建新的object，其中键来自一个object，值来自另一个

Question

I have the following input:

{
  "Columns": [
    {
      "email": 123,
      "name": 456,
      "firstName": 789,
      "lastName": 450,
      "admin": 900,
      "licensedSheetCreator": 617,
      "groupAdmin": 354,
      "resourceViewer": 804,
      "id": 730,
      "status": 523,
      "sheetCount": 298
    }
  ]
}
{
  "Users": [
    {
      "email": "abc@def.com",
      "name": "Abc Def",
      "firstName": "Abc",
      "lastName": "Def",
      "admin": false,
      "licensedSheetCreator": true,
      "groupAdmin": false,
      "resourceViewer": true,
      "id": 521,
      "status": "ACTIVE",
      "sheetCount": 0
    },
    {
      "email": "aaa@bbb.com",
      "name": "Aaa Bob",
      "firstName": "Aaa",
      "lastName": "Bob",
      "admin": false,
      "licensedSheetCreator": true,
      "groupAdmin": false,
      "resourceViewer": false,
      "id": 352,
      "status": "ACTIVE",
      "sheetCount": 0
    }
  ]
}

I need to change the key for all key value pairs in users to match the value in Columns, like so:

{
  "Columns": [
    {
      "email": 123,
      "name": 456,
      "firstName": 789,
      "lastName": 450,
      "admin": 900,
      "licensedSheetCreator": 617,
      "groupAdmin": 354,
      "resourceViewer": 804,
      "id": 730,
      "status": 523,
      "sheetCount": 298
    }
  ]
}
{
  "Users": [
    {
      123: "abc@def.com",
      456: "Abc Def",
      789: "Abc",
      450: "Def",
      900: false,
      617: true,
      354: false,
      804: true,
      730: 521,
      523: "ACTIVE",
      298: 0
    },
    {
      123: "aaa@bbb.com",
      456: "Aaa Bob",
      789: "Aaa",
      450: "Bob",
      900: false,
      617: true,
      354: false,
      804: false,
      730: 352,
      523: "ACTIVE",
      298: 0
    }
  ]
}

I don't mind if I update the Users array or create a new array of objects. I have tried several combinations of with entries, to entries, from entries, trying to search for keys using variables but the more I dive into it, the more confused I get.

Answer 1

Elements of a stream are processed independently. So we have to change the input.

We could group the stream elements into an array. For an input stream, this can be achieved using --slurp / -s . ^[1]

jq -s '
   ( .[0].Columns[0] | map_values( tostring ) ) as $map |
   (
      .[0],
      (
         .[1:][] |
         .Users[] |= with_entries(
            .key = $map[ .key ]
         )
      )
   )
'

Demo on jqplay

Alternatively, we could use --null-input / -n in conjunction with input and/or inputs to read the input.

jq -n '
   input |
   ( .Columns[0] | map_values( tostring ) ) as $map |
   (
      .,
      (
         inputs |
         .Users[] |= with_entries(
            .key = $map[ .key ]
         )
      )
   )
'

Demo on jqplay

Note that your desired output isn't valid JSON. Object keys must be strings. So the above produces a slightly different document than requested.

Note that I assumed that .Columns is always an array of one exactly one element. This is a nonsense assumption, but it's the only way the question makes sense.

---

1. For a stream the code generates, you could place the stream generator in an array constructor ( [] ). reduce can also be used to collect from a stream. For example, map(... ) can be written as [.[] |... ] and as reduce.[] as $_ ( []; . + [ $_ |... ] ) .

Answer 2

The following has the merit of simplicity, though it does not sort the keys. It assumes jq is invoked with the -n option and of course produces a stream of valid JSON objects:

input
| . as $Columns
| .Columns[0] as $dict
| input # Users
| .Users[] |= with_entries(.key |= ($dict[.]|tostring))
| $Columns, .

If having the keys sorted is important, then you could easily add suitable code to do that; alternatively, if you don't mind having the keys of all objects sorted, you could use the -S command-line option.