Batch analyse is slow and seems to not generate PDF with OCR

Question

Batch analyse is slow and seems to not generate PDF with OCR

Pedro Oliveira 0

Hi,

We are trying to use Document Intelligence Batch Analyze in our document OCR workflow.

But we were unable to use the PDF generation feature on this option we tried the following C# code so that it would be created.

string MODEL_ID = "prebuilt-read";
AnalyzeBatchDocumentsOptions options = new AnalyzeBatchDocumentsOptions(
            MODEL_ID,
            new BlobContentSource(sourceContainerUri)
            {
                // Prefix = sourceBlobNamePrefix
            },
            destinationContainerUri
        );

        // options.ResultPrefix = destinationBlobNamePrefix;
        options.Features.Add(DocumentAnalysisFeature.Barcodes);
        
        // cant make this work
        options.Output.Add(AnalyzeOutputOption.Pdf);
        options.OutputContentFormat = new DocumentContentFormat("pdf");

        var result = await _client.AnalyzeBatchDocumentsAsync(
            WaitUntil.Started,
            options,
            cancellationToken
        );

On the configured storage we saw no PDF generated with the OCR results. We also tried looking inside the json files, but there was no PDF binary there. Where can we get the PDF?

Also, it takes some time, in the order of minutes, to handle the batch. When picking up a single file to handle it is quite fast but between the files it's slow. How can we configure so that the batch is always handled correctly?

Regards, Pedro Oliveira

Manas Mohanty 8,150 Reputation points Microsoft External Staff Moderator

2025-07-30T06:31:24.95+00:00

Hi Pedro Oliveira

There were few outages earlier in document intelligence Does the issue still replicate at your side

Could you please share full code to replicate the issue.

https://github.com/Azure/azure-sdk-for-python/blob/main/sdk/documentintelligence/azure-ai-documentintelligence/samples/sample_analyze_batch_documents.py

Attached relevant document that I am referring to replicate the issue.

Thank you
Manas Mohanty 8,150 Reputation points Microsoft External Staff Moderator

2025-08-04T03:21:30.9733333+00:00

Hi Pedro Oliveira

I had tried to output into pdf format with output content format option. It did not work as you mentioned.

We need to use external python library to save the results with OCR detection in a blank pdf file.

https://learn.microsoft.com/en-us/answers/questions/2153111/how-to-batch-create-a-searchable-pdf-using-azure-d

Thank you

1 answer

Your answer

Manas Mohanty 8,150 Reputation points Microsoft External Staff Moderator

2025-07-30T06:31:24.95+00:00

Hi Pedro Oliveira

There were few outages earlier in document intelligence Does the issue still replicate at your side

Could you please share full code to replicate the issue.

https://github.com/Azure/azure-sdk-for-python/blob/main/sdk/documentintelligence/azure-ai-documentintelligence/samples/sample_analyze_batch_documents.py

Attached relevant document that I am referring to replicate the issue.

Thank you
Manas Mohanty 8,150 Reputation points Microsoft External Staff Moderator

2025-08-04T03:21:30.9733333+00:00

Hi Pedro Oliveira

I had tried to output into pdf format with output content format option. It did not work as you mentioned.

We need to use external python library to save the results with OCR detection in a blank pdf file.

https://learn.microsoft.com/en-us/answers/questions/2153111/how-to-batch-create-a-searchable-pdf-using-azure-d

Thank you

Answer 1

Hi Pedro,

Thank you for contacting Q&A Forum.

To address the issue you're facing with Document Intelligence Batch Analyze, here are some steps you can follow:

Getting Batch Results: Follow the instructions provided in the Azure documentation to retrieve the batch results. https://learn.microsoft.com/en-us/dotnet/api/azure.ai.documentintelligence.analyzebatchresult?view=azure-dotnet-preview
Checking Storage Configuration: If the batch jobs were successful but no result file can be found in the target container, I would suggest you check for the storage container configuration to verify that the Document Intelligence resource has the necessary permissions to write files to this container.
Handling Failed Batch Jobs: If the batch jobs failed, follow the error messages to identify the issues and make the necessary adjustments before retrying the batch jobs.

For more detailed information on batch jobs, you can refer to the Azure documentation on batch analysis. https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/prebuilt/batch-analysis?view=doc-intel-4.0.0

If I have answered your question, please accept this answer as a token of appreciation and don't forget to give a thumbs up for "Was it helpful"!

Best Regards,

Pedro Oliveira 0 Reputation points

2025-07-01T15:39:35.27+00:00

Hi Dang,

Our problem isn´t the results, those were written to the container.

The problem is that we can´t find the generated PDF, that we requested with the option AnalyzeOutputOption.Pdf.

Also the batches don´t fail, but it takes too long to process the documents, in the order of minutes between picking up another document.

Kind regards,

Pedro Oliveira

Share via

Batch analyse is slow and seems to not generate PDF with OCR

1 answer

Your answer