CODEWITHSUNDEEP
jq
Haim Marko
Haim Marko

Reputation: 55

JQ duplicate results in output json

I have the following input json for jq:

{
  "records": [
    {
      "uuid": "14e20b5a-6619-11ec-a838-0050568168f7",
      "nas": {
        "path": "/data/nlp/datasets/ds0"
      }
    },
    {
      "uuid": "799ce816-656c-11ec-a838-0050568168f7",
      "nas": {
        "path": "/data/nlp/datasets/ds1"
      }
    }
  ]
}

What is the reason the following jq duplicate the results:

[{junction:.records[].nas.path}+{uuid:.records[].uuid}]

Result:

[
  {
    "junction": "/data/nlp/datasets/ds0",
    "uuid": "14e20b5a-6619-11ec-a838-0050568168f7"
  },
  {
    "junction": "/data/nlp/datasets/ds1",
    "uuid": "14e20b5a-6619-11ec-a838-0050568168f7"
  },
  {
    "junction": "/data/nlp/datasets/ds0",
    "uuid": "799ce816-656c-11ec-a838-0050568168f7"
  },
  {
    "junction": "/data/nlp/datasets/ds1",
    "uuid": "799ce816-656c-11ec-a838-0050568168f7"
  }
]

Upvotes: 1

Views: 1024

Answers (1)

pmf
pmf

Reputation: 36033

Both {junction:.records[].nas.path}and {uuid:.records[].uuid} produce a stream of two objects each due to the iteration caused by .records[]. Adding up the cartesian product, this totals to four objects.

If you want to iterate jut once, pull out the iterator .records[] up front to define a common input context for the subsequent filter:

.records[] | {junction:.nas.path} + {uuid:.uuid}

This can then further be simplified to

.records[] | {junction:.nas.path, uuid}

{
  "junction": "/data/nlp/datasets/ds0",
  "uuid": "14e20b5a-6619-11ec-a838-0050568168f7"
}
{
  "junction": "/data/nlp/datasets/ds1",
  "uuid": "799ce816-656c-11ec-a838-0050568168f7"
}

which is presumably what you were looking for.

Demo

Upvotes: 3

Related Questions