How Does a Search Index Handle Updated PDF Files with the Same Name but Different Content?

[svenfeld]

How does a search index behave when a PDF file with the same name but different content is ingested? Specifically:

Does the search index automatically remove the chunks that reference the old version of the PDF? Or does it simply add new chunks without removing the old ones?

In this scenario, is it better to delete the old document from the index before ingesting the new version?

[pamelafox]

For this repo, the manual ingestion assumes that the same name is the same file. Our manual ingestion script does not attempt to remove any chunks, you would need to manually remove them if desired.
See:
https://github.com/Azure-Samples/azure-search-openai-demo/blob/main/docs/data_ingestion.md#removing-documents

However, perhaps you're asking about how Azure AI search indexers work? There is more details about that in the documentation here:
https://learn.microsoft.com/en-us/azure/search/search-howto-index-changed-deleted-blobs?tabs=portal