How to delete a document from Search index? Any suggestions how to do it

[rehat22]

I used the following script to delete a dcument from azure ai search:

import requests service_name = "" index_name = "" api_version = "2023-07-01-preview" admin_api_key = "" document_key = ""

url = f"https://{service_name}.search.windows.net/indexes/{index_name}/docs/index?api-version={api_version}"

data = { "value": [{ "id": document_key, "@search.action": "delete" }] }

headers = { "Content-type": "application/json", "api-key": admin_api_key }

response = requests.post(url, headers=headers, json=data)

if response.status_code == 200 or response.status_code == 204: print("Document deleted successfully!") else: print(f"Error deleting document: {response.status_code} {response.text}").

I got a success message on running it but the document is still there in the index. Is there an another way to approach this?

[guyyardeni]

The chat interaction uses the chunk index as well, so you'd have to clear the chunks to get the data out of the index. I'm not sure if there is a way to identify which chunks are associated with a specific file so I delete all of the chunk and reset the indexers so they create new chunks and index them.

[rehat22]

thank you @guyyardeni

[jack-vinitsky]

The chat interaction uses the chunk index as well, so you'd have to clear the chunks to get the data out of the index. I'm not sure if there is a way to identify which chunks are associated with a specific file so I delete all of the chunk and reset the indexers so they create new chunks and index them.

You should be able to use the parent_id field located in the chunk index. This is the id of the original document. All chunks derived from a given a document will have the same value for this field.

[jack-vinitsky]

The chat interaction uses the chunk index as well, so you'd have to clear the chunks to get the data out of the index. I'm not sure if there is a way to identify which chunks are associated with a specific file so I delete all of the chunk and reset the indexers so they create new chunks and index them.

You should be able to use the parent_id field located in the chunk index. This is the id of the original document. All chunks derived from a given a document will have the same value for this field.

@rehat22 If this answer resolves your issue, please mark it as closed.

[rehat22]

Hi, You can do it. Thank you

[jack-vinitsky]

Hi, You can do it. Thank you

I think only the person who opened the issue or an admin can do it.