如何将JSON格式的oracle表行值转换为多个关系COLUMNS？

Question

I have a oracle table that has been loaded by MQ and one of the table column has values with JSON format data.I need to convert those JSON format data into relational row format. I am wondering if there is any SQL in oracle with/without creating temp tables to convert those JSON format data into relational column values?

Select JSON_MG from test_1;

JSON_MG
{
  "type": "testeevnet",
  "version": "test-1.0",
  "testsource": "1.0.0",
  "timestamp": "2019-02-14T20:45:18.4422201+00:00",
  "test_ID": 11,
  "tag": "22",
  "PNAME": "test/test_n",
  "ticket": "WT9999",
  "ticketStatus": "active",
  "tickets": [
    {
      "ticket": "convert",
      "code": "PA",
      "date": "2019-03-31"
    }
    }
  ]
}

-- Expected result all the JSON elements as columns

Answer 1

json_table is what you're looking for. This can do JSON -> relational conversion.

Just list out the paths to the elements you want to extract & their data type:

with jdata as (
  select '{
  "type": "testeevnet",
  "version": "test-1.0",
  "testsource": "1.0.0",
  "timestamp": "2019-02-14T20:45:18.4422201+00:00",
  "test_ID": 11,
  "tag": "22",
  "PNAME": "test/test_n",
  "ticket": "WT9999",
  "ticketStatus": "active",
  "tickets": [
    {
      "ticket": "convert",
      "code": "PA",
      "date": "2019-03-31"
    }
  ]
}' doc
  from dual
)
  select t.* 
  from   jdata, json_table (
    doc, '$'
    columns (
      type    varchar2 path '$.type',
      version varchar2 path '$.version',
      nested  path '$.tickets[*]' 
      columns (
        ticket varchar2 path '$.ticket'
      )
    )
  ) t;

TYPE         VERSION    TICKET    
testeevnet   test-1.0   convert   

The nested path splits out the array elements into rows. So as you add more objects in your tickets array, they become rows:

with jdata as (
  select '{
  "type": "testeevnet",
  "version": "test-1.0",
  "testsource": "1.0.0",
  "timestamp": "2019-02-14T20:45:18.4422201+00:00",
  "test_ID": 11,
  "tag": "22",
  "PNAME": "test/test_n",
  "ticket": "WT9999",
  "ticketStatus": "active",
  "tickets": [
    {
      "ticket": "convert",
      "code": "PA",
      "date": "2019-03-31"
    },
    {
      "ticket": "convert2",
      "code": "PA",
      "date": "2019-03-31"
    }
  ]
}' doc
  from dual
)
  select t.* 
  from   jdata, json_table (
    doc, '$'
    columns (
      type    varchar2 path '$.type',
      version varchar2 path '$.version',
      nested  path '$.tickets[*]' 
      columns (
        ticket varchar2 path '$.ticket'
      )
    )
  ) t;

TYPE         VERSION    TICKET     
testeevnet   test-1.0   convert     
testeevnet   test-1.0   convert2