ネストされたjsonの複数のレイヤーをSQLテーブルに変換する方法

json_normalize リストが必要です 辞書の場合とEbsの場合 -これは単なる辞書なので、JSONデータを前処理する必要があります：

In [88]: with open(fn) as f:
    ...:     data = json.load(f)
    ...:

In [89]: for r in data['Volumes']:
    ...:     if 'Ebs' not in r: # add 'Ebs' dict if it's not in the record...
    ...:         r['Ebs'] = []
    ...:     if not isinstance(r['Ebs'], list): # wrap 'Ebs' in a list if it's not a list 
    ...:         r['Ebs'] = [r['Ebs']]
    ...:

In [90]: data
Out[90]:
{'Volumes': [{'Attachments': [{'AttachTime': '2013-12-18T22:35:00.000Z',
     'DeleteOnTermination': True,
     'Device': '/dev/sda1',
     'InstanceId': 'i-1234567890abcdef0',
     'State': 'attached',
     'Tags': [{'Key': 'Name', 'Value': 'DBJanitor-Private'},
      {'Key': 'Owner', 'Value': 'DBJanitor'},
      {'Key': 'Product', 'Value': 'Database'},
      {'Key': 'Portfolio', 'Value': 'DB Janitor'},
      {'Key': 'Service', 'Value': 'DB Service'}],
     'VolumeId': 'vol-049df61146c4d7901'}],
   'AvailabilityZone': 'us-east-1a',
   'Ebs': [{'AttachTime': '2016-09-14T19:49:11.000Z',
     'DeleteOnTermination': True,
     'Status': 'attached',
     'VolumeId': 'vol-049df61146c4d7901'}],
   'VolumeId': 'vol-049df61146c4d7901',
   'VolumeType': 'standard'}]}

注：'Ebs': {..} 'Ebs': [{..}]

In [91]: e = pd.io.json.json_normalize(data['Volumes'],
    ...:                               ['Ebs'],
    ...:                               ['VolumeId'],
    ...:                               meta_prefix='parent_')
    ...:


In [92]: e
Out[92]:
                 AttachTime DeleteOnTermination    Status               VolumeId        parent_VolumeId
0  2016-09-14T19:49:11.000Z                True  attached  vol-049df61146c4d7901  vol-049df61146c4d7901