将 Json 转换为 SQL 表

Posted

技术标签:

【中文标题】将 Json 转换为 SQL 表【英文标题】：Converting Json to SQL table 【发布时间】：2017-03-19 22:27:02 【问题描述】：

我正在尝试学习如何将以下格式的 json 转换为 sql 表。我使用了 python pandas，它正在将 json 节点转换为字典。

相同的json：


    "Volumes": [
        
            "AvailabilityZone": "us-east-1a",
            "Attachments": [
                
                    "AttachTime": "2013-12-18T22:35:00.000Z",
                    "InstanceId": "i-1234567890abcdef0",
                    "VolumeId": "vol-049df61146c4d7901",
                    "State": "attached",
                    "DeleteOnTermination": true,
                    "Device": "/dev/sda1"
                
            ],
            "Tags": [
            
                "Value": "DBJanitor-Private",
                "Key": "Name"
            ,
            
                "Value": "DBJanitor",
                "Key": "Owner"
            ,
            
                "Value": "Database",
                "Key": "Product"
            ,
            
                "Value": "DB Janitor",
                "Key": "Portfolio"
            ,
            
                "Value": "DB Service",
                "Key": "Service"
            
        ],
            "VolumeType": "standard",
            "VolumeId": "vol-049df61146c4d7901",
            "State": "in-use",
            "SnapshotId": "snap-1234567890abcdef0",
            "CreateTime": "2013-12-18T22:35:00.084Z",
            "Size": 8
        ,
        
            "AvailabilityZone": "us-east-1a",
            "Attachments": [],
            "VolumeType": "io1",
            "VolumeId": "vol-1234567890abcdef0",
            "State": "available",
            "Iops": 1000,
            "SnapshotId": null,
            "CreateTime": "2014-02-27T00:02:41.791Z",
            "Size": 100
        
    ]

直到现在..这是我正在尝试的......在python中：

asg_list_json_Tags=asg_list_json["AutoScalingGroups"]
Tags=pandas.DataFrame(asg_list_json_Tags)
n = []
for i in Tags.columns:
    n.append(i)
print n

engine = create_engine("mysql+mysqldb://user:"+'pwd'+"@mysqlserver/dbname")
Tags.to_sql(name='TableName', con=engine, if_exists='append', index=True)

【问题讨论】：

似乎是什么问题？为什么该代码不起作用？ 所以我得到一个错误说 dict 不能插入到字符串中 @DataJanitor，你想存储 flatten 数据吗？ @MaxU - 是的！这就是我想做的事 @DataJanitor，问题是你想用Attachments做什么？有些记录是缺失的，所以这里不能使用json_normalize，因为没有Attachments的记录不会被解析... 【参考方案1】：

我会这样做：

fn = r'D:\temp\.data\40450591.json'

with open(fn) as f:
    data = json.load(f)

# some of your records seem NOT to have `Tags` key, hence `KeyError: 'Tags'`
# let's fix it
for r in data['Volumes']:
    if 'Tags' not in r:
        r['Tags'] = []

v = pd.DataFrame(data['Volumes']).drop(['Attachments', 'Tags'],1)
a = pd.io.json.json_normalize(data['Volumes'], 'Attachments', ['VolumeId'], meta_prefix='parent_')
t = pd.io.json.json_normalize(data['Volumes'], 'Tags', ['VolumeId'], meta_prefix='parent_')

v.to_sql('volume', engine)
a.to_sql('attachment', engine)
t.to_sql('tag', engine)

输出：

In [179]: v
Out[179]:
                      AvailabilityZone                CreateTime    Iops  Size              SnapshotId      State VolumeType
VolumeId
vol-049df61146c4d7901       us-east-1a  2013-12-18T22:35:00.084Z     NaN     8  snap-1234567890abcdef0     in-use   standard
vol-1234567890abcdef0       us-east-1a  2014-02-27T00:02:41.791Z  1000.0   100                    None  available        io1

In [180]: a
Out[180]:
                 AttachTime DeleteOnTermination     Device           InstanceId     State               VolumeId        parent_VolumeId
0  2013-12-18T22:35:00.000Z                True  /dev/sda1  i-1234567890abcdef0  attached  vol-049df61146c4d7901  vol-049df61146c4d7901
1  2013-12-18T22:35:11.000Z                True  /dev/sda1  i-1234567890abcdef1  attached  vol-049df61146c4d7111  vol-049df61146c4d7901

In [217]: t
Out[217]:
         Key              Value        parent_VolumeId
0       Name  DBJanitor-Private  vol-049df61146c4d7901
1      Owner          DBJanitor  vol-049df61146c4d7901
2    Product           Database  vol-049df61146c4d7901
3  Portfolio         DB Janitor  vol-049df61146c4d7901
4    Service         DB Service  vol-049df61146c4d7901

测试 JSON 文件：


    "Volumes": [
        
            "AvailabilityZone": "us-east-1a",
            "Attachments": [
                
                    "AttachTime": "2013-12-18T22:35:00.000Z",
                    "InstanceId": "i-1234567890abcdef0",
                    "VolumeId": "vol-049df61146c4d7901",
                    "State": "attached",
                    "DeleteOnTermination": true,
                    "Device": "/dev/sda1"
                ,
                
                    "AttachTime": "2013-12-18T22:35:11.000Z",
                    "InstanceId": "i-1234567890abcdef1",
                    "VolumeId": "vol-049df61146c4d7111",
                    "State": "attached",
                    "DeleteOnTermination": true,
                    "Device": "/dev/sda1"
                
            ],
            "Tags": [
                
                    "Value": "DBJanitor-Private",
                    "Key": "Name"
                ,
                
                    "Value": "DBJanitor",
                    "Key": "Owner"
                ,
                
                    "Value": "Database",
                    "Key": "Product"
                ,
                
                    "Value": "DB Janitor",
                    "Key": "Portfolio"
                ,
                
                    "Value": "DB Service",
                    "Key": "Service"
                
            ],
            "VolumeType": "standard",
            "VolumeId": "vol-049df61146c4d7901",
            "State": "in-use",
            "SnapshotId": "snap-1234567890abcdef0",
            "CreateTime": "2013-12-18T22:35:00.084Z",
            "Size": 8
        ,
        
            "AvailabilityZone": "us-east-1a",
            "Attachments": [],
            "VolumeType": "io1",
            "VolumeId": "vol-1234567890abcdef0",
            "State": "available",
            "Iops": 1000,
            "SnapshotId": null,
            "CreateTime": "2014-02-27T00:02:41.791Z",
            "Size": 100
        
    ]

【讨论】：

谢谢@MaxU 真的很感激！如果有任何问题，我会将其插入并给您发送错误 打败我！使用匹配的 ID 捕获一对多以进行数据库规范化的好方法。 @DataJanitor，当然，很高兴我能帮上忙 @MaxU - 我能够成功实现代码！如果我还有一层复杂性，我该如何解决？对问题中的 json 进行了更改。请看问题 @DataJanitor，你可以像Attachments一样解析它。注意：key Tags 必须在每条记录中出现，否则将无法正常工作。每个卷记录中都有Tags 吗？【参考方案2】：

类似于此示例：https://github.com/zolekode/json-to-tables/blob/master/example.py

使用以下脚本：

以下脚本将数据导出为 html，但您也可以将其导出为 SQL。

table_maker.save_tables(YOUR_PATH, export_as="sql", sql_connection=YOUR_CONNECTION)
# See the code below

import json
from extent_table import ExtentTable
from table_maker import TableMaker

Volumes = [
    
        "AvailabilityZone": "us-east-1a",
        "Attachments": [
            
                "AttachTime": "2013-12-18T22:35:00.000Z",
                "InstanceId": "i-1234567890abcdef0",
                "VolumeId": "vol-049df61146c4d7901",
                "State": "attached",
                "DeleteOnTermination": "true",
                "Device": "/dev/sda1"
            
        ],
        "Tags": [
            
                "Value": "DBJanitor-Private",
                "Key": "Name"
            ,
            
                "Value": "DBJanitor",
                "Key": "Owner"
            ,
            
                "Value": "Database",
                "Key": "Product"
            ,
            
                "Value": "DB Janitor",
                "Key": "Portfolio"
            ,
            
                "Value": "DB Service",
                "Key": "Service"
            
        ],
        "VolumeType": "standard",
        "VolumeId": "vol-049df61146c4d7901",
        "State": "in-use",
        "SnapshotId": "snap-1234567890abcdef0",
        "CreateTime": "2013-12-18T22:35:00.084Z",
        "Size": 8
    ,
    
        "AvailabilityZone": "us-east-1a",
        "Attachments": [],
        "VolumeType": "io1",
        "VolumeId": "vol-1234567890abcdef0",
        "State": "available",
        "Iops": 1000,
        "SnapshotId": "null",
        "CreateTime": "2014-02-27T00:02:41.791Z",
        "Size": 100
    
]

volumes = json.dumps(Volumes)
volumes = json.loads(volumes)

extent_table = ExtentTable()
table_maker = TableMaker(extent_table)
table_maker.convert_json_objects_to_tables(volumes, "volumes")
table_maker.show_tables(8)
table_maker.save_tables("./", export_as="html") # you can also pass in export_as="sql" or "csv". In the case of sql, there is a parameter to pass the engine.

HTML 输出：

<table border="1" class="dataframe">
  <thead>
    <tr style="text-align: right;">
      <th>ID</th>
      <th>AvailabilityZone</th>
      <th>VolumeType</th>
      <th>VolumeId</th>
      <th>State</th>
      <th>SnapshotId</th>
      <th>CreateTime</th>
      <th>Size</th>
      <th>Iops</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>0</td>
      <td>us-east-1a</td>
      <td>standard</td>
      <td>vol-049df61146c4d7901</td>
      <td>in-use</td>
      <td>snap-1234567890abcdef0</td>
      <td>2013-12-18T22:35:00.084Z</td>
      <td>8</td>
      <td>None</td>
    </tr>
    <tr>
      <td>1</td>
      <td>us-east-1a</td>
      <td>io1</td>
      <td>vol-1234567890abcdef0</td>
      <td>available</td>
      <td>null</td>
      <td>2014-02-27T00:02:41.791Z</td>
      <td>100</td>
      <td>1000</td>
    </tr>
    <tr>
      <td>2</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
    </tr>
  </tbody>
</table>
<table border="1" class="dataframe">
  <thead>
    <tr style="text-align: right;">
      <th>ID</th>
      <th>PARENT_ID</th>
      <th>is_scalar</th>
      <th>scalar</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>0</td>
      <td>0</td>
      <td>False</td>
      <td>None</td>
    </tr>
  </tbody>
</table>

<table border="1" class="dataframe">
  <thead>
    <tr style="text-align: right;">
      <th>ID</th>
      <th>AttachTime</th>
      <th>InstanceId</th>
      <th>VolumeId</th>
      <th>State</th>
      <th>DeleteOnTermination</th>
      <th>Device</th>
      <th>PARENT_ID</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>0</td>
      <td>2013-12-18T22:35:00.000Z</td>
      <td>i-1234567890abcdef0</td>
      <td>vol-049df61146c4d7901</td>
      <td>attached</td>
      <td>true</td>
      <td>/dev/sda1</td>
      <td>0</td>
    </tr>
    <tr>
      <td>1</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
    </tr>
  </tbody>
</table>

<table border="1" class="dataframe">
  <thead>
    <tr style="text-align: right;">
      <th>ID</th>
      <th>PARENT_ID</th>
      <th>is_scalar</th>
      <th>scalar</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>0</td>
      <td>0</td>
      <td>False</td>
      <td>None</td>
    </tr>
    <tr>
      <td>1</td>
      <td>0</td>
      <td>False</td>
      <td>None</td>
    </tr>
    <tr>
      <td>2</td>
      <td>0</td>
      <td>False</td>
      <td>None</td>
    </tr>
    <tr>
      <td>3</td>
      <td>0</td>
      <td>False</td>
      <td>None</td>
    </tr>
    <tr>
      <td>4</td>
      <td>0</td>
      <td>False</td>
      <td>None</td>
    </tr>
  </tbody>
</table>

<table border="1" class="dataframe">
  <thead>
    <tr style="text-align: right;">
      <th>ID</th>
      <th>Value</th>
      <th>Key</th>
      <th>PARENT_ID</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td>0</td>
      <td>DBJanitor-Private</td>
      <td>Name</td>
      <td>0</td>
    </tr>
    <tr>
      <td>1</td>
      <td>DBJanitor</td>
      <td>Owner</td>
      <td>1</td>
    </tr>
    <tr>
      <td>2</td>
      <td>Database</td>
      <td>Product</td>
      <td>2</td>
    </tr>
    <tr>
      <td>3</td>
      <td>DB Janitor</td>
      <td>Portfolio</td>
      <td>3</td>
    </tr>
    <tr>
      <td>4</td>
      <td>DB Service</td>
      <td>Service</td>
      <td>4</td>
    </tr>
    <tr>
      <td>5</td>
      <td>None</td>
      <td>None</td>
      <td>None</td>
    </tr>
  </tbody>
</table>

【讨论】：

问题是如何转换为 SQL，而不是 HTML 表。 @szeta 这有点不公平，你不觉得吗？你看到上面这行代码了吗？ table_maker.save_tables("./", export_as="html") # you can also pass in export_as="sql" or "csv". In the case of sql, there is a parameter to pass the engine. 我说导出可以是从html到sql的任何东西。 HTML 表格输出是为了提供视觉效果。我希望你重新考虑你的投票。 我没有看到这条评论。为避免将来发生这种情况，我建议更好地突出这些要点。回答这个问题最相关的一点是（是！）隐藏在评论深处，远非显而易见。如果您编辑答案，我将撤销反对票。 （现在在编辑之前被阻止）

以上是关于将 Json 转换为 SQL 表的主要内容，如果未能解决你的问题，请参考以下文章

将sql表数据转换为json

将表转换为像字典一样的 json 对象 SQL？

将 SQL 结果转换为 JSON 的正确方法

PHP 将 SQL 请求转换为 JSON

python：sql建表语句转换为json

前端json转sql