How to extract multiple nested attributes from all items in a list of dicts?

I have a list of objects and I need to pull out multiple attributes from the items in the list so that the result is a list of only those attributes and none of the others. Here's what my original data looks like:

[
  {
    "failed": true,
    "msg": "Error while parsing document: files/file1.xml (Opening and ending tag mismatch)",
    "invocation": {
      "module_args": {
        "path": "files/file1.xml"
      }
    }
  },
  {
    "failed": true,
    "msg": "Error while parsing document: files/file2.xml (Opening and ending tag mismatch)",
    "invocation": {
      "module_args": {
        "path": "files/file2.xml"
      }
    }
  }
]

I need to map the msg and invocation.module_args.path attributes into new objects. The resulting list should look like this:

[
  {
    "msg": "Error while parsing document: files/file1.xml (Opening and ending tag mismatch)",
    "path": "files/file1.xml"
  },
  {
    "msg": "Error while parsing document: files/file2.xml (Opening and ending tag mismatch)",
    "path": "files/file2.xml"
  }
]

If I can possibly define new attribute names that would be preferred, but it's lower priority:

[
  {
    "error": "Error while parsing document: files/file1.xml (Opening and ending tag mismatch)",
    "file": "files/file1.xml"
  },
  {
    "error": "Error while parsing document: files/file2.xml (Opening and ending tag mismatch)",
    "file": "files/file2.xml"
  }
]

I can do this by creating an additional task file and looping over it, creating each dict object and updating the list but I'd rather just use a single task and/or Jinja2 filter if possible to simplify things.

Is this possible?

Solution

Given the mre data

    l1:
      - failed: true
        invocation:
          module_args: {path: files/file1.xml}
        msg: 'Error: files/file1.xml'
      - failed: true
        invocation:
          module_args: {path: files/file2.xml}
        msg: 'Error: files/file2.xml'

Use json_query. For example,

  err: "{{ l1|json_query('[].{error: msg,
                              file: invocation.module_args.path}') }}"

gives what you want

  err:
  - error: 'Error: files/file1.xml'
    file: files/file1.xml
  - error: 'Error: files/file2.xml'
    file: files/file2.xml

^{Example of a complete playbook for testing
- hosts: all

vars:

l1:
- failed: true
invocation:
module_args: {path: files/file1.xml}
msg: 'Error: files/file1.xml'
- failed: true
invocation:
module_args: {path: files/file2.xml}
msg: 'Error: files/file2.xml'

err: "{{ l1|json_query('[].{error: msg,
file: invocation.module_args.path}') }}"
tasks:

- debug:
var: err}