jq: translate array of objects to object

[EDIT: The following response was in answer to the question when it described (a) the mapping as shown below, and (b) the input data as having the form:

[  {    "list": [      {        "value": 1,        "id1": 12      },      {        "value": 15,        "id2": 13      },      {        "value": -4,        "id3": 14      }    ]  }]

END OF EDIT]

In the following I'll assume that the mapping is available via the following function, but that is an inessential assumption:

def mapping: {  "id1": "newId1",  "id2": "newId2",  "id3": "newId3"} ;

The following jq filter will then produce the desired output:

map( .list     |= (map( to_entries[]              | (mapping[.key]) as $mapped              | select($mapped)              | {($mapped|tostring): .value} )         | add) )

There's plenty of ways to skin a cat. I'd do it like this:

.[].list |= reduce .[] as $i ({};    ($i.id|tostring) as $k      | (select($mapping | has($k))[$mapping[$k]] = $i.value) // .)

You would just provide the mapping through a separate file or argument.

$ cat program.jq.[].list |= reduce .[] as $i ({};    ($i.id|tostring) as $k      | (select($mapping | has($k))[$mapping[$k]] = $i.value) // .)$ cat mapping.json{  "12": "newId1",  "13": "newId2",  "14": "newId3"}$ jq --argfile mapping mapping.json -f program.jq input.json[  {    "list": {      "newId1": 1,      "newId2": 15,      "newId3": -4    }  }]

Here is a reduce-free solution to the revised problem.

In the following I'll assume that the mapping is available via the following function, but that is an inessential assumption:

def mapping:{  "12": "newId1",  "13": "newId2",  "14": "newId3"} ;map( .list     |= (map( mapping[.id|tostring] as $mapped              | select($mapped)              |  {($mapped): .value} )         | add) )

The "select" is for safety (i.e., it checks that the .id under consideration is indeed mapped). It might also be appropriate to ensure that $mapped is a string by writing {($mapped|tostring): .value}.