如何避免Graphene在查询响应时对OrderedDict进行排序？（Python，Django）

huangapple
go评论159阅读模式
英文:

How to avoid Graphene sorting an OrderedDict on query response? (Python, Django)

问题

我试图从Django应用程序中的Graphene端点返回一个OrderedDict的字典，使用了GenericScalar类型字段。

OrderedDict根据值排序，但API似乎重新按键对字典进行排序。

OrderedDict的项目格式为：{ id : [name, amount] }

我发送的OrderedDicts之一的示例：

  1. { 'foods': {
  2.     3: ['apple', 5],
  3.     1: ['banana', 7],
  4.     2: ['carrot', 3]
  5.   }, 
  6.   'animals': {
  7.     5: ['cat', 3],
  8.     3: ['dog', 10],
  9.     1: ['pig', 5],
  10.   }
  11. }

在端点响应中收到的内容：

  1. { 'foods': {
  2.     1: ['banana', 7],
  3.     2: ['carrot', 3],
  4.     3: ['apple', 5],
  5.   }, 
  6.   'animals': {
  7.     1: ['pig', 5],
  8.     3: ['dog', 10],
  9.     5: ['cat', 3],
  10.   }
  11. }

我特别使用OrderedDicts，因为顺序很重要，但我找不到一种方法来避免Graphene按键重新排序OrderedDicts。

这是ObjectType和Query的声明方式，尽管没有什么特别之处：

  1. class DashboardType(g.ObjectType):
  2.     data = GenericScalar()
  4. class Query(object):
  5.     dashboard = Field(DashboardDataType, id=ID(), date=Date())
  6.    
  7.     def resolve_dashboard(self, info, **kwargs):
  8.         data = function_to_get_the_data(kwargs)
  9.         # 到这一步为止，数据仍然按预期排序
  10.         return data

数据按预期排序，直到Graphene API发送响应。我期望顺序能够保持，但实际情况并非如此。

有关如何实现此目标的建议吗？
我没想到会这么困难，但我已经没有主意了。

英文:

I am trying to return a dict of OrderedDicts from a Graphene endpoint in a Django application, by using the GenericScalar type field.
The OrderedDicts are sorted based on value, but the API seems to re-sort the dicts by key.

The OrderedDicts items format is: { id : [name, amount] }

Example of one of the OrderedDicts that I send:
<!-- language: python -->

  1. { &#39;foods&#39;: {
  2.     3: [&#39;apple&#39;, 5],
  3.     1: [&#39;banana&#39;, 7],
  4.     2: [&#39;carrot&#39;, 3]
  5.   }, 
  6.   &#39;animals&#39;: {
  7.     5: [&#39;cat&#39;, 3],
  8.     3: [&#39;dog&#39;, 10],
  9.     1: [&#39;pig&#39;, 5],
  10.   }
  11. }

What is received in the endpoint response:
<!-- language: python -->

  1. { &#39;foods&#39;: {
  2.     1: [&#39;banana&#39;, 7],
  3.     2: [&#39;carrot&#39;, 3],
  4.     3: [&#39;apple&#39;, 5],
  5.   }, 
  6.   &#39;animals&#39;: {
  7.     1: [&#39;pig&#39;, 5],
  8.     3: [&#39;dog&#39;, 10],
  9.     5: [&#39;cat&#39;, 3],
  10.   }
  11. }

I am specifically using OrderedDicts because the order is important, but I have not found a way of avoiding Graphene to re-sort the OrderedDicts by key.

This is the how the ObjectType and Query are declared, althought is is nothing out of the ordinary.
<!-- language: python -->

  1. class DashboardType(g.ObjectType):
  2.     data = GenericScalar()
  4. class Query(object):
  5.     dashboard = Field(DashboardDataType, id=ID(), date=Date())
  6.    
  7.     def resolve_dashboard(self, info, **kwargs):
  8.         data = function_to_get_the_data(kwargs)
  9.         # up until this point the data is still sorted correctly
  10.         return data

The data is sorted as expected up until the response is sent by the Graphene API. I would expect the order to be mantained, but it is not.

Any suggestions on how to achieve this?
I didn't think it would be so difficult, but I am out of ideas.

答案1

得分: 0

  1. 我建议一个模型：
  3. ```graphql
  4. type 东西 {
  5.   id: ID!
  6.   名字: String!
  7.   子东西: [子东西]
  8. }
  10. type 子东西 {
  11.   id: ID!
  12.   名字: String!
  13.   数量: Int
  14. }

然后你可以有一个查询：

  1. 查询 获取东西(按顺序: String, 方向: String): [东西]

然后你可以执行：

  1. 查询 按名字获取东西($按顺序: '名字', $方向: '升序') {
  2.   查询 获取东西(按顺序: $按顺序, 方向: $方向) {
  3.     id
  4.     名字
  5.     子东西 {
  6.       id
  7.       名字
  8.       数量
  9.     }
  10.   }
  11. }

然后你的 子东西 解析器可以根据 按顺序 和 方向 查询参数对 子东西 进行排序。

然后你的返回对象将如下所示：

  1. [
  2.   {
  3.     "id": `ID`,
  4.     "名字": "食物",
  5.     "子东西": [
  6.       {
  7.         "id": 3,
  8.         "名字": "苹果",
  9.         "数量": 5
  10.       },
  11.       等等…
  12.     ]
  13.   }
  14. ]
英文:

I would suggest as a model:

  1. type Thing {
  2.   id: ID!
  3.   name: String!
  4.   subthings: [Subthing]
  5. }
  7. type Subthing {
  8.   id: ID!
  9.   name: String!
  10.   amount: Int
  11. }

Then you could have a guery:

  1. query getThings(orderBy: String, direction: String): [Thing]

And against that you could execute:

  1. query getThingsOrderByName($orderBy: &#39;name&#39;, $direction: &#39;asc&#39;) {
  2.   query getThings(orderBy: $orderBy, direction: $direction) {
  3.     id
  4.     name
  5.     subthings {
  6.       id
  7.       name
  8.       amount
  9.     }
  10.   }
  11. }

Your subThings resolver could then sort the subThings based on the orderBy and 'direction' query parameters.

Your return object would then look like:

  1. [
  2.   {
  3.     &quot;id&quot;: `
  4.     &quot;name&quot;: &quot;foods&quot;
  5.     &quot;subthings&quot;: [
  6.       {
  7.         &quot;id&quot;: 3,
  8.         &quot;name&quot;: &quot;apple&quot;,
  9.         &quot;amount&quot;: 5,
  10.       },
  11.       etc…
  12.     ]
  13.   }
  14. ]

huangapple
  • 本文由 发表于 2023年7月13日 23:58:53
  • 转载请务必保留本文链接：https://go.coder-hub.com/76681328.html
  • django
  • graphene-python
  • graphql
  • python-3.x
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定