Convert table in nested JSON

[spaz]

What is the easiest way to convert a table to a nested JSON?

[AmanSehgal]

@Sergio Paz​ could you please provide an example of your input table and expected output JSON file?

[spaz]

{

  "accounts":[

     "12516898"

  ],

  "deals":[

     {

        "dealId":"4897143",

        "promotionId":"AVC84897143",

        "conditions":{

           "paymentMethod":null,

           "simulationDateTime":{

              "startDateTime":"2022-03-16",

              "endDateTime":"2022-04-30"

           },

           "scaledLineItem":{

              "sharedMinimumQuantity":true,

              "skus":[

                 "000000000000000031"

              ],

              "crossDiscount":true,

              "ranges":[

                 {

                    "index":"1",

                    "from":1,

                    "to":200

                 }

              ]

           }

        },

        "output":{

           "lineItemScaledDiscount":{

              "ranges":[

                 {

                    "index":"1",

                    "skus":[

                       "000000000000000031"

                    ],

                    "type":"%",

                    "discount":5.5,

                    "maxQuantity":null,

                    "proportion":null,

                    "fixed":true

                 }

              ]

           }

        }

     }

  ]

}

[-werners- · Accepted solution]

As Spark can handle nested columns, I would first construct the nested structure in spark (as from spark 3.1.1 there is the excellent column.withField method with which you can create your structure.

Finally write it to json.

That seems to be the easiest way, but your case might be more complex, that is hard to say without some more info.

[Anonymous]

@Sergio Paz​ - How's it going? Are you able to give us more information?

[jose_gonzalez]

Hi @Sergio Paz​ ,

Just a friendly follow-up. Could you provide more details on your use case? please share your code snippet, so we can help you.