From the GraphAPI I receive the following (pre-filtered) response, however there are still values that I don't want. I want to keep it as simple as possible, but not sure which approach is the best here. One of the approaches I tried is posted down here.

{
  "@odata.context": "https://graph.microsoft.com/v1.0/$metadata#users('[email protected]')/messages(sender,subject,body)",
  "value": [
    {
      "@odata.etag": "W/\"6xAAAAAAKq\"",
      "id": "6xAAAAAAVQAAA=",
      "subject": "Can we speak tomorrow?",
      "body": {
        "contentType": "text",
        "content": "Meeting at 10?"
      },
      "sender": {
        "emailAddress": {
          "name": "test-mail",
          "address": "[email protected]"
        }
      }
    },
    {
      "@odata.etag": "W/\"6xAAAAAAKq\"",
      "id": "6xAAAAAAVQAAA=",
      "subject": "Meet for lunch?",
      "body": {
        "contentType": "text",
        "content": "The new cafeteria is open."
      },
      "sender": {
        "emailAddress": {
          "name": "test-mail",
          "address": "[email protected]"
        }
      }
    },
    {
      "@odata.etag": "W/\"6xAAAAAAKq\"",
      "id": "6xAAAAAAVQAAA=",
      "subject": "Nice city",
      "body": {
        "contentType": "text",
        "content": "Want to join me on a business trip?"
      },
      "sender": {
        "emailAddress": {
          "name": "test-mail",
          "address": "[email protected]"
        }
      }
    }
  ]
}

The result should look like this:

{
  "Messages": [
    {
      "id": "6xAAAAAAVQAAA=",
      "sender": "[email protected]",
      "subject": "Can we speak tomorrow?",
      "content": "Meeting at 10?"
    },
    {
      "id": "6xAAAAAAVQAAA=",
      "sender": "[email protected]",
      "subject": "Meet for lunch?",
      "content": "The new cafeteria is open."
    },
    {
      "id": "6xAAAAAAVQAAA=",
      "sender": "[email protected]",
      "subject": "Nice city",
      "content": "Want to join me on a business trip?"
    }
  ]
}

I have tried the following, but can't really get my head around it.

[
  {
    "operation": "shift",
    "spec": {
      "value": {
        "*": {
          "id": "value.[&1].id",
          "sender": "value.[&1].sender.emailAddress.address",
          "subject": "value.[&1].subject",
          "body": "value.[&1].sender.emailAddress.address"
        }
      }
    }
  }
]

CodePudding user response：

It's a good idea to walk through the indices of the value list within a shift transformation spec. You can prefer addressing the attributes, which are not in the current object, by @(0,elm_at_surr_level.sub_elm_level1.sub_elm_level2) reference as key such as

[
  {
    "operation": "shift",
    "spec": {
      "value": {
        "*": {
          "id": "Messages.[&1].&",
          "@(0,sender.emailAddress.address)": "Messages.[&1].sender",
          "subject": "Messages.[&1].&",
          "@(0,body.content)": "Messages.[&1].content"
        }
      }
    }
  }
]

the demo on the site

If sorting as displayed(in the order of id->sender->subject->content) needed, then that would be a bit tricky such as

[
  {
    "operation": "shift",
    "spec": {
      "value": {
        "*": {
          "id": "Messages.[&1].&",
          "@(0,sender.emailAddress.address)": "Messages.[&1].sender",
          "subject": "Messages.[&1].&",
          "@(0,body.content)": "Messages.[&1].zontent"
        }
      }
    }
  },
  {
    "operation": "sort"
  },
  {
    "operation": "shift",
    "spec": {
      "Messages": {
        "*": {
          "*": "&2.[&1].&",
          "z*": "&2.[&1].c&(0,1)"
        }
      }
    }
  }
]