How to create seperate tables from postgreSQL JSON column value

Question

Following is the table format :

CREATE TABLE records
(
  id text,
  record json
) 

For every record column of the row , JSON value will be: (this is just a snapshot of an entire structure)

{
    "Trial": {
        "primary_id": "xxxxx",
        "key1": "aaaaaaaa (BAL-8557) aaaaaaaaaaaaaaaaaa",
        "key2": "aaaaaaaaaaaaa , aaaaaaaaaaaaaaaaaaa",
        "key3": "Yes",
        "key4": "No",
        "key5": {
            "key5k": [{
                    "@type": "ABC",
                    "$": "ABC121 "
                },
                {
                    "@type": "ABC",
                    "$": "ABC12133 "
                }
            ]
        },
        "Indications": {
            "Indication": [{
                    "indication_id": "1308",
                    "$": "aaaaaa"
                },
                {
                    "indicationid": "1309",
                    "$": "bbbbbb"
                }
            ]
        }
    }
}

there are around 300K rows in the table currently

I want to develop an utility that will read all rows find keys which have nested JSON.

Based on the nested JSON structure I want to create a separate table

Eg from above JSON table , Indication Table will be

----------------------------------------
primary_id  |   indication_id |   value
----------------------------------------
    xxxxx       |   1308      | aaaaaa
----------------------------------------
    xxxxx       |   1309      | bbbbbb

---

How can I develop this utility ?

Should this be handled by PostgreSQL functions or using Python + Spark ?

If someone could share any link for this solution or something similar that will be helpful , thanks