无法使用适用于 Azure AI 搜索的 Python SDK 将数据添加到 ComplexField

我想将带有嵌套字典的有效负载上传到 Azure AI 搜索索引。我在索引中使用 ComplexField 作为负载中的嵌套字典。索引无法识别嵌套字典，并且出现空错误。这是我的代码：

    ComplexField,
    CorsOptions,
    SearchIndex,
    ScoringProfile,
    SearchFieldDataType,
    SimpleField,
    SearchableField,
)

request_example = {
        
            "id": "1",
            "text": "bla",
            "payload": {
                "processId": "00665",
                "color": "green"
            }
        
    }

# Create a search index  
index_client = SearchIndexClient(endpoint=service_endpoint, credential=credential)  
fields = [  
    SimpleField(name="id", type=SearchFieldDataType.String, key=True, sortable=True, filterable=True, facetable=True),  
    SearchableField(name="text", type=SearchFieldDataType.String),  
    ComplexField(
        name="payload",
        fields=[
            SimpleField(name="processId", type=SearchFieldDataType.String),
            SimpleField(name="color", type=SearchFieldDataType.String),
        ],
        collection=True,
    )
]  
  
# Configure the vector search configuration  
vector_search = VectorSearch(  
    algorithms=[  
        HnswAlgorithmConfiguration(  
            name="myHnsw",  
            kind=VectorSearchAlgorithmKind.HNSW,  
            parameters=HnswParameters(  
                m=4,  
                ef_construction=400,  
                ef_search=500,  
                metric=VectorSearchAlgorithmMetric.COSINE,  
            )  
        ),  
    ],  
    profiles=[  
        VectorSearchProfile(  
            name="myHnswProfile",  
            algorithm_configuration_name="myHnsw",   
        ),  
    ],  
)  
  
semantic_config = SemanticConfiguration(  
    name="my-semantic-config",  
    prioritized_fields=SemanticPrioritizedFields(  
        # title_field=SemanticField(field_name="title"),  
        # keywords_fields=[SemanticField(field_name="color")],  
        content_fields=[SemanticField(field_name="text")]  
    )  
)  
  
# Create the semantic settings with the configuration  
semantic_search = SemanticSearch(configurations=[semantic_config])  
cors_options = CorsOptions(allowed_origins=["*"], max_age_in_seconds=60)
from typing import List
scoring_profiles: List[ScoringProfile] = []
# Create the search index with the semantic settings  
index = SearchIndex(name="complex_field_test", fields=fields,  
                    vector_search=vector_search, semantic_search=semantic_search,scoring_profiles=scoring_profiles, cors_options=cors_options)  
result = index_client.create_or_update_index(index)  
print(f'{result.name} created')  

# Upload some documents to the index
search_client = SearchClient(endpoint=service_endpoint, index_name="complex_field_test", credential=credential)
result = search_client.upload_documents([request_example])  
print(f"Uploaded {len(documents)} documents")

我得到的错误：

The request is invalid. Details: A null value was found for the property named 'payload', which has the expected type 'Collection(search.complex.payload)[Nullable=False]'. The expected type 'Collection(search.complex.payload)[Nullable=False]' does not allow null values.

因此，据我了解，嵌套字典“有效负载”未被识别并且显示为空，尽管子字段已在索引中注册并且没有空值。你能帮我解决这个问题吗？如何向 ComplexField 添加数据？

在使用 Azure AI 搜索的 Python SDK 时遇到的错误信息表明，的索引定义预期“payload”字段是一个复杂类型数组 (Collection(search.complex.payload))，但它收到的值是单个对象。

为了解决这个问题，需要将“payload”的值包装在一个列表中，即使它只有一个元素：

request_example = {
    "id": "1",
    "text": "bla",
    "payload": [  # Wrap the payload object in a list
        {
            "processId": "00665",
            "color": "green"
        }
    ]
}

通过将“payload”的值包装在一个列表中，告诉 Azure AI 搜索这是一个复杂类型数组，这与的索引定义相匹配。

另外，请确保的 ComplexField 定义中 collection=True 已经设置。这将确保该字段可以接受多个复杂对象。

以下是更正后的代码片段：

# ... other code ...

# Upload some documents to the index
search_client = SearchClient(endpoint=service_endpoint, index_name="complex_field_test", credential=credential)
result = search_client.upload_documents([request_example])  # "payload" is now a list
print(f"Uploaded {len(result)} documents") 

通过进行此更改，应该能够成功将数据上传到包含 ComplexField 的 Azure AI 搜索索引。