Modify/Create new azure ai search index with "IMPORT AND VECTORIZE DATA" option and configure some fields as facetable, sortable

Question

Modify/Create new azure ai search index with "IMPORT AND VECTORIZE DATA" option and configure some fields as facetable, sortable

Rikin 20

Hey Azure Community,

I want to ask you a question in regards of azure ai search. Actually I am not able to make any fields facetable, sortable, etc when I try to create indexes using "IMPORT AND VECTORIZE DATA" option.

Can you please guide me how to make some fields facetable, sortable while creating a new indexes using "IMPORT AND VECTORIZE DATA" option.

If I try to create index using "IMPORT DATA" option I am able to configure fields but I do not want to use traditional (text) based search index. I want to make index which uses semantic+vector search.

Thanks.

Siva Nair 1,475 Reputation points Microsoft External Staff

2025-04-05T03:21:41.68+00:00

Hi Rikin,

Just checking back to see if the above comment helped or you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.
Siva Nair 1,475 Reputation points Microsoft External Staff

2025-04-06T12:17:51.04+00:00

Hi Rikin,

Just checking back to see if the above comment helped or you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Accepted answer

0 additional answers

Your answer

Siva Nair 1,475 Reputation points Microsoft External Staff

2025-04-05T03:21:41.68+00:00

Hi Rikin,

Just checking back to see if the above comment helped or you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.
Siva Nair 1,475 Reputation points Microsoft External Staff

2025-04-06T12:17:51.04+00:00

Hi Rikin,

Just checking back to see if the above comment helped or you have a resolution yet. In case if you have any resolution, please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

Answer 1

Siva Nair 1,475 Microsoft External Staff

Hi Rikin,

The "Import and Vectorize Data" option auto-generates fields optimized for AI search and doesn't allow customization. Manually defining the schema before indexing gives you full control over facetable, sortable, and filterable attributes.

To achieve it we should manually create the index and then link it to a data source.

Navigate to Azure AI Search → Indexes → "Create Index" and define your index schema manually. When configuring the fields, ensure that you set the necessary attributes: facetable as true for filtering and grouping, sortable as true for ordering results, filterable as true for using in $filter queries, and retrievable as true if you want the field to appear in search results. Additionally, for vector search, define a field of type Collection(Edm.Single) and attach a vectorSearchConfiguration.

Once the index is created, you need to manually attach a data source (such as Azure Blob Storage, Cosmos DB, or SQL Database) and configure an indexer to populate the index. Unlike the automated process in "Import and Vectorize Data," this approach allows you to fine-tune the schema while still benefiting from semantic+vector search.

{
  "name": "my-search-index",
  "fields": [
    {
      "name": "id",
      "type": "Edm.String",
      "key": true,
      "retrievable": true
    },
    {
      "name": "title",
      "type": "Edm.String",
      "retrievable": true,
      "sortable": true,
      "facetable": true
    },
    {
      "name": "category",
      "type": "Edm.String",
      "filterable": true,
      "facetable": true
    },
    {
      "name": "contentVector",
      "type": "Collection(Edm.Single)",
      "searchable": true,
      "vectorSearchConfiguration": "default"
    }
  ],
  "vectorSearch": {
    "algorithms": [
      {
        "name": "default",
        "kind": "hnsw",
        "parameters": {
          "m": 4,
          "efConstruction": 400
        }
      }
    ]
  }
}

If you have any further assistant, do let me know.

Rikin 20 Reputation points

2025-04-07T15:02:35.1466667+00:00

Hey Siva,

Thanks for answering my query but if possible can you please also lay down the steps of how to manually attach a data source (such as Azure Blob Storage, Cosmos DB, or SQL Database) and configure an indexer to populate the index.

Thanks Again.

Best,
Rikin
Siva Nair 1,475 Reputation points Microsoft External Staff

2025-04-07T15:54:51.9733333+00:00
Hi Rikin,

Sure! to manually connect a data source like Azure Blob Storage, Cosmos DB, or SQL Database to Azure AI Search, and configure an indexer to populate a search index, lets follow these steps:

Navigate to your Azure AI Search service. If you don't have one, create a new search service in your desired region and pricing tier.

Attach your data source. Go to the "Data sources" blade in your search service and click on "+ Add". Choose the type of data source: Blob Storage, Cosmos DB, or SQL Database.

For Azure Blob Storage, enter a name for the data source, choose your storage connection method (connection string or Managed Identity), and select the blob container you want to index. Optionally, you can provide a folder path or blob prefix to scope the data. You can also set up a data change detection policy (e.g., using LastModified timestamps) and a deletion detection policy.

For Azure Cosmos DB, enter the data source name and provide the connection string or use Managed Identity. Select the database and container (collection), and optionally provide a custom query to filter data. It’s recommended to set a change detection policy using a HighWaterMarkColumnName like _ts.

For Azure SQL Database, enter a name and the connection string (again, with username/password or Managed Identity). Specify the table or view you want to index. You may also provide a custom SQL query and a column like ModifiedDate as the high-water mark for change detection.

After setting up the data source, go to the "Indexes" blade and click "+ Add" to create a new search index. Give your index a name and define its schema — this includes the field names, data types, and attributes like searchable, filterable, sortable, etc. Be sure to select a key field (typically an ID) which uniquely identifies each document in the index. You can manually define the schema or use the "Import Data" wizard to generate it from your source structure.

Then, move to the "Indexers" section and click "+ Add". Name the indexer, and link it to the previously created data source and index. Optionally, configure field mappings if the field names in your source and index differ. You can also define how often the indexer should run (e.g., every 5 minutes, hourly, daily) by setting up a schedule.

Once the indexer is created, you can run it manually from the portal to populate the index immediately. The status and logs of the indexer runs can be monitored from the "Execution history" section to ensure data is being ingested as expected.

Finally, to verify your index, go to Search Explorer within the Azure AI Search service. Select your index and run a sample query like { "search": "*", "top": 5 } to view the first few results.

Please refer, https://learn.microsoft.com/en-us/azure/search/search-what-is-data-import

https://learn.microsoft.com/en-us/azure/search/search-how-to-create-search-index?tabs=portal

https://learn.microsoft.com/en-us/azure/search/search-howto-indexing-azure-blob-storage

If you have any further assistant, do let me know.

Rikin 20

Hey Siva,

I tried the approach you mentioned and it worked fine but when I checked my Vector index quota usage which is 0 Bytes. Which I think shouldn't be zero. can you please let me know how to resolve this: User's image

I also have my vector defined in my indexes schema as below:

"vectorSearch": {

"algorithms": [

  {

    "name": "rikin-indexer-algorithm",

    "kind": "hnsw",

    "hnswParameters": {

      "metric": "cosine",

      "m": 4,

      "efConstruction": 400,

      "efSearch": 500

    }

  }

],

"profiles": [

  {

    "name": "rikin-indexer-azureOpenAi-text-profile",

    "algorithm": "rikin-indexer-algorithm",

    "vectorizer": "rikin-indexer-azureOpenAi-text-vectorizer"

  }

],

"vectorizers": [

  {

    "name": "rikin-indexer-azureOpenAi-text-vectorizer",

    "kind": "azureOpenAI",

    "azureOpenAIParameters": {

      "resourceUri": "https://open-ai.openai.azure.com",

      "deploymentId": "text-embedding-ada-002",

      "apiKey": "<redacted>",

      "modelName": "text-embedding-ada-002"

    }

  }

],

Can you please answer my query. Thanks.

Siva Nair 1,475 Microsoft External Staff

Hi Rikin,

Can you once test weather vectors are present or not following this test-a-vectorizer.

### Run a query
 POST {{baseUrl}}/indexes/vector-nasa-ebook-txt/docs/search?api-version=2024-07-01 HTTP/1.1
     Content-Type: application/json
     api-key: {{queryApiKey}}
     {
         "count": true,
         "select": "title,chunk",
         "vectorQueries": [
             {
                 "kind": "text",
                 "text": "what cloud formations exists in the troposphere",
                 "fields": "vector",
                 "k": 3,
                 "exhaustive": true
             }
         ]
     }

Let me know the test result.

Share via

Modify/Create new azure ai search index with "IMPORT AND VECTORIZE DATA" option and configure some fields as facetable, sortable

0 additional answers

Your answer