Parsing JSON response from API using DataFlow

How do I parse a JSON response from API as CSV?
Currently the API response is stored as a JSON file on Blob Storage and I now want to parse and save it as CSV file.

Here's a subset of the JSON response as sample,

{
  "reports": [
    {
      "columnHeader": {
        "dimensions": [
          "ga:dimension2",
          "ga:dimension1",
          "ga:dimension4",
          "ga:dimension6"
        ],
        "metricHeader": {
          "metricHeaderEntries": [
            {
              "name": "ga:newUsers",
              "type": "INTEGER"
            }
          ]
        }
      },
      "data": {
        "rows": [
          {
            "dimensions": [
              "Australia",
              "Hong Kong",
              "HKNG",
              "not set"
            ],
            "metrics": [
              {
                "values": [
                  "1"
                ]
              }
            ]
          },
          {
            "dimensions": [
              "Australia",
              "Malaysia",
              "KL",
              "not set"
            ],
            "metrics": [
              {
                "values": [
                  "1"
                ]
              }
            ]
          }
        ]
      }
    }
  ]
}

the desired output.csv file should have column headers from the data key of JSON file.

the column header would ideally be like this,
+-----------+------------+------------+------------+-------------+
|dimension1 | dimension2 | dimension4 | dimension6 | metric_value|
+-----------+------------+------------+------------+-------------+
| Australia | Hong Kong  | HKNG       | not set    | 1           |
| ...       | ...        | ....       | ...        | ...         |
+-----------+------------+------------+------------+-------------+

To parse a JSON response from an API and save it as a CSV file in Azure Data Factory (ADF) Data Flow, you can use the following steps:

• Take the Source transformation in dataflow and take the JSON file from Blob Storage as the source dataset. Set the documentForm option to arrayOfDocuments. This will indicate that the JSON file contains an array of objects.

• Take the Select transformation and map the data field of the second object in the reports array to a new column called data.

data = reports[1].data

• Take the Flatten transformation and unroll the values array in the metrics field of the data column. Map the dimensions and values fields ato new columns.
  

• Then take the Derive transformation to create new columns based on the dimensions and values columns. New columns called dimensions2, dimensions1, dimensions4, dimensions6, and metric_values are created here.

• Take the Select transformation is to select dimensions2, dimensions1, dimensions4, dimensions6, and metric_values fields only.

• Then take the sink transformation with csv dataset.

Output:

When the pipeline with this dataflow is run, the Json data will be copied to csv file in required format.