Azure data factory data flow json to SQL

Question

I have a source JSON file with hierarchical data, which I need to sink in two SQL tables(relational). The JSON is as below

{
"orders":[
  {
    "orderid":"30933",
    "ordername":"abc",
    "items":[
      {
        "itemid":1,
        "itemstatus":"Failed"
      },
      {
        "itemid":2,
        "itemstatus":"Failed"
      }
    ]
  },
  {
    "orderid":"308320",
    "ordername":"xyz",
    "items":[
      {
        "itemid":5,
        "itemstatus":"Succeeded"
      }
    ]
  }
]
}

My SQL holding two tables Order and OrderItem with OrderID primary and foreign key.

Now I have an Azure data factory data flow with source as above JSON and I need to park all data relational in respective tables.

So here I need OrderId(30933,308320) and OrderName(abc,xyz) will go into Order table and respective items data go into OrderItem table(which reference OrderId from Order table). In this case Order table have 2 and OrderItem table have 3 entries.

Answer 1

We can not achieve that in one copy active.

We could using two copy actives in one pipeline, I tested and it succeed. You could follow my steps bellow:

1. Copy active1: copy the data from Orders (orderid and ordername) to table Orders .
2. Copy active2: copy the data from items (itemid and itemstatus) to table OrderItems .

Note:

1. Copy active 1 and 2 use the same json file as the source. The differences are in the Mapping settings.

2. Copy active 1 sink is Azure SQL database table Orders , Copy active 2 sink is Azure SQL database table OrderItems .

To make you understand it clearly, I made two GIF pictures.

Mapping settings in Copy active 1:

Mapping settings in Copy active 2:

Run the pipeline:

Check the data in table:

The limit is that we only could get the first element of items , we can not choose the collection reference both.

Update:

Congratulations that you achieved it in another way: flatten/transpose the data using data flow and then maintain it to pour in a relational SQL table.