My search results on my index return no results

Question

In the Azure AI Foundry Chat Playground, I used the "Add a new data source" feature to create a new vector index. When creating the vector index, I chose to upload a pdf file as my data source.

However, once the vector index is created and I use the chat to ask about information within the pdf file, I get the following error as seen in the screenshot above:

"Error

400: An error occurred when calling Azure Cognitive Search: Azure Search: Please assign a proper column/field for vector search. It should be of type Collection(Edm.Single)"

Why do I get this error, and how do I resolve this?

Additionally, when I go to "Data + indexes" and try create a new index here using the same pdf file that I upload again, I get an error when clicking on the "Create vector index" button at the end of the process. The error states as follows:

"Failed to create vector index. PermissionDenied: Principal does not have access to API/Operation. Trace ID : b1028240-c122-4e60-8833-c2b5b2b0f4feClient request ID : 8f9a5ac3-1aea-4b96-9567-757f6625821bAPIM request ID : 86c9adda-43c2-4cec-9e61-936b11dabf40"

When I open my Azure Search Service, the index that I created in the Azure AI Foundry Chat Playground is seen there. However, if I use the search explorer in the search service to query anything in the vector index, I get no results (as seen in my first screenshot attached).

Is the cause of these three issues the same thing? If so, how can I fix this?

If there are several issues, please help resolve them all.

Answer 1

Hi @Nathanael de Swardt ,
Thanks for reaching out to Microsoft Q&A.
To fix Azure Cognitive Search not returning results and vector search errors, first check your index schema and ensure the vector field is correctly defined. A common error, "Please assign a proper column/field for vector search. It should be of type Collection(Edm.Single)", usually means the vector field is missing or not set right. Go to your Azure Cognitive Search service, open the index, and verify if the vector field exists and is set as Collection(Edm.Single). Update the schema if needed. Supported Data Types in Azure Search for correct field setup.

If you see the "PermissionDenied: Principal does not have access to API/Operation" error when creating a vector index, it means you lack permissions. Head to Azure Cognitive Search → Access Control (IAM) → Role Assignments, and ensure your user, service principal, or managed identity has the Cognitive Services Contributor or Cognitive Search Contributor role. Add permissions if needed, then try again. More details are at Azure Q&A Permissions Error.

If your index returns no results when queried via the Azure Search Explorer, it might be due to failed document processing or improper data indexing. Go to Azure Cognitive Search → Indexes → Your Index, check the document count, and ensure records are there. If no documents are found, re-upload your data using the Azure portal, SDKs, or REST APIs, and set an appropriate embedding model when setting up the vector index. Azure AI Search Indexing.

After indexing, make sure queries are correctly formatted. In the Azure Search Explorer, test a broad query like { "search": "*" } to check for results. If no results are found, confirm your AI Search index is connected to the Azure AI Foundry chat. A query syntax guide is available at Azure Search Query Syntax. If the issue persists, check logs in Azure Cognitive Search → Monitoring → Diagnostic Settings to identify potential problems.
If the answer is helpful, please click Accept Answer and kindly upvote it so that other people who faces similar issue may get benefitted from it.
Please let us know if you need further assistance.