Storing large strings in dynamoDB about fhir-works-on-aws

[ssvegaraju]

Hi @nkundella,

thanks for raising this issue. Have you considered leveraging Binary resources to store these pdf files? We store the FHIR Resource for Binary Objects in DDB, but put the large binary data within S3. Would this work for your use case? You could then fill the extensions with references to these Binary resources.

Thanks for your time,
Sukeerth

[nkundella]

I did think of saving jsons to s3 and using the s3 urls in the extensions to refer back to the pdf files.
But in my case, if there is a case where we want to export the json and give to an organization. We have to get the pdf back from s3 and attach it to the extension and then send the complete json.
Would that be possible? or is there any other way?

[ssvegaraju]

When using Binary resources in FWoA, performing a GET request will return a presigned S3 GET URL from which you can download the pdf from S3, and then attach it back to the json. If you want to avoid using Binary resources and instead store directly in S3, you could construct something similar and use a middleware to get the resource from S3 and attach it to the JSON. However, both of these approaches may increase latency time in requests as there is download time from S3 involved. Does this answer your question?