CODEWITHSUNDEEP
jsonlinuxgroup-byjq
developthou
developthou

Reputation: 363

How to group by in jq?

Here's the json document

[
    {"name": "bucket1","clusterName":"cluster1"},
    {"name": "bucket2","clusterName":"cluster1"},
    {"name": "bucket3","clusterName":"cluster2"},
    {"name": "bucket4","clusterName":"cluster2"}
]

And I want to convert it to

[
{"clusterName": "cluster1", buckets:[{"name": "bucket1"}, {"name": "bucket2"}]},
{"clusterName": "cluster2", buckets:[{"name": "bucket1"}, {"name": "bucket2"}]},
]

How do I do that in jq?

Upvotes: 5

Views: 2977

Answers (3)

mgaert
mgaert

Reputation: 2388

Using map and based on peak's answer, I needed all the bucket names in an array:

jq 'group_by(.clusterName) | map( { clusterName : .[0].clusterName, buckets: [ .[].name ] } )'

gives:

[
  {
    "clusterName": "cluster1",
    "buckets": [
      "bucket1",
      "bucket2"
    ]
  },
  {
    "clusterName": "cluster2",
    "buckets": [
      "bucket3",
      "bucket4"
    ]
  }
]

Upvotes: 0

peak
peak

Reputation: 116670

Using map makes for quite a tidy answer:

group_by(.clusterName)
| map( {clusterName: .[0].clusterName,
        buckets: map( {name} ) } )

Upvotes: 1

Torben Pi Jensen
Torben Pi Jensen

Reputation: 870

cat doc.json | jq '[group_by(.clusterName)[] | { clusterName: .[0].clusterName, "buckets": [.[] | { name: .name } ]}]'

Should do what you want. Now with an array around as well.

Upvotes: 3

Related Questions