Integrated vectorization - sourcefile and storageUrl is null

[poonkuzhali]

For data ingestion, I wanted to enable integrated vectorization. I followed the instructions provided in the docs -> data_ingestion.md file

1. I deleted the old index
2. Enabled azd env set USE_FEATURE_INT_VECTORIZATION true
3. azd provision
4. azd deploy
5. An indexer, index and data source were created.

All my documents are present in blob storage.

After the indexer successfully ran, I used the app, the citation url was wrong, it wasnt linked to the file in blob storage.

When I looked at the index, the sourcefile and storageUrl variables were null, which is messing with my citations.

The citation does not have an extension, just has the file name. I am not sure why

[poonkuzhali]

@pamelafox Any thoughts on this issue would be much appreciated. Thank you

[timwillittes]

I can help with the root cause analysis. If you go to the integrated vectorization file (app/backend/prepdocslib/integratedvectorizerstrategy.py), on lines 100-102 you will see where the index field mappings are taking place. The sourcefile and storageUrl attributes are not mapped, and I'm not sure if that is on purpose or not.

selectors=[
                SearchIndexerIndexProjectionSelector(
                    target_index_name=index_name,
                    parent_key_field_name="parent_id",
                    source_context="/document/pages/*",
                    mappings=[
                        InputFieldMappingEntry(name="content", source="/document/pages/*"),
                        InputFieldMappingEntry(name="embedding", source="/document/pages/*/vector"),
                        InputFieldMappingEntry(name="sourcepage", source="/document/metadata_storage_name"),
                    ],
                ),
            ],

[cforce]

InputFieldMappingEntry(name="sourcefile", source="/document/metadata_storage_path"),
InputFieldMappingEntry(name="storageUrl", source="/document/metadata_storage_url"),                    ],

??

[wvdm1217]

Hi, I am also having the same issue.

Tried mapping sourcefile and storageurl in the skills section but it seems to not be working. Not sure where to view the necessary metadata

    "indexProjections": {
        "selectors": [
            {
                "targetIndexName": "gptkbindex",
                "parentKeyFieldName": "parent_id",
                "sourceContext": "/document/pages/*",
                "mappings": [
                    {
                        "name": "content",
                        "source": "/document/pages/*",
                        "inputs": []
                    },
                    {
                        "name": "embedding",
                        "source": "/document/pages/*/vector",
                        "inputs": []
                    },
                    {
                        "name": "sourcepage",
                        "source": "/document/metadata_storage_name",
                        "inputs": []
                    },
                    {
                        "name": "sourcefile",
                        "source": "/document/metadata_storage_path",
                        "inputs": []
                    },
                    {
                        "name": "storageUrl",
                        "source": "/document/metadata_storage_url",
                        "inputs": []
                    }
                ]
            }
        ],
        "parameters": {
            "projectionMode": "skipIndexingParentDocuments"
        }
    }

[pamelafox]

I am closing this issue since we are deprecating the previous integrated vectorization feature.
We have made a custom skillset that runs our own code instead, mirroring the local ingestion in Azure Functions. We recommend the new skillset approach for maximal compatibility with the repo features, including page numbers.

Please try it out and raise any issues you run into. Thanks!

https://github.com/Azure-Samples/azure-search-openai-demo/releases/tag/2025-11-12