如何使用jq打印指定对象键？优化冗余代码实现目标输出

Problem #

I'm working with AWS Elasticsearch domain JSON data and already have a working jq command to get my desired Option 2 output, but it's redundant because I keep repeating the .ElasticsearchClusterConfig path. I want to streamline this using a cleaner syntax like .ElasticsearchClusterConfig | {InstanceType, InstanceCount} instead.

Here's the simplified structure of my input JSON:

{ 
  "DomainStatus": { 
    "DomainId": "12345", 
    "DomainName": "test", 
    "ARN": "arn:aws:es:eu-west-1:12345:domain/test", 
    "Created": true, 
    "Deleted": false, 
    "Endpoint": "test.com", 
    "Processing": false, 
    "UpgradeProcessing": false, 
    "ElasticsearchVersion": "5.3", 
    "ElasticsearchClusterConfig": { 
      "InstanceType": "t2.medium.elasticsearch", 
      "InstanceCount": 2, 
      "DedicatedMasterEnabled": false, 
      "ZoneAwarenessEnabled": true, 
      "ZoneAwarenessConfig": { "AvailabilityZoneCount": 2 } 
    }, 
    "EBSOptions": { "EBSEnabled": true, "VolumeType": "gp2", "VolumeSize": 30 } 
  } 
}

My current working (but redundant) command for Option 2:

jq '.DomainStatus | {DomainName, Endpoint, InstanceType: .ElasticsearchClusterConfig.InstanceType, InstanceCount: .ElasticsearchClusterConfig.InstanceCount}'

This gives me:

{
  "DomainName": "test",
  "Endpoint": "test.com",
  "InstanceType": "t2.medium.elasticsearch",
  "InstanceCount": 2
}

Solution #

You can use jq's object merge operator (+) to combine two separate objects—one with your top-level fields, and another with the extracted cluster config fields—without repeating the path reference.

Optimized Command for Option 2

jq '.DomainStatus | {DomainName, Endpoint} + (.ElasticsearchClusterConfig | {InstanceType, InstanceCount})'

Breakdown of how this works:

1. First, we create an object with just DomainName and Endpoint directly from .DomainStatus
2. Next, we extract InstanceType and InstanceCount from .ElasticsearchClusterConfig into a second object
3. The + operator merges these two objects into a single one, giving you exactly the flat structure you want for Option 2

Bonus: Command for Option 1 (Nested Cluster Config)

If you ever need to go back to Option 1's nested structure, you can cleanly nest the extracted cluster fields like this:

jq '.DomainStatus | {DomainName, Endpoint, ElasticsearchClusterConfig: (.ElasticsearchClusterConfig | {InstanceType, InstanceCount})}'

This keeps your code concise, readable, and free of redundant path references.

内容的提问来源于stack exchange，提问作者mrk