How big is the PDF (file size)? Also, it seems weird that you have to take the detour through base64, the BLOB contents should already have the binary file contents, unencoded.

One of the test PDFs is ~21.1MB and most will be around this size, 10MB~30MB. I have tried the BLOB directly with many different content-types, but it created a corrupt PDF each time. This experience led me to this other community thread: How to extract PDF attachment from ArcGIS Online and upload to online location with HTTPCaller? (safe.com) and that got me to where I am with it just almost working. The original data or encoded to Base 64 is able to produce a valid and complete PDF locally using a FileWriter, but not the SharePoint API being used here.

There are several things that could be the problem. What I'm saying about the BLOB, is that if you save it directly (as a binary file) with the file extension .pdf, it should open just fine in e.g. Acrobat. In other words, the BLOB should contain exactly the file contents on the file system.

If you're holding a 30MB file in an attribute and then do a Base64 encoding, you're suddenly holding a 40MB (~33% increase for Base64) attribute. I'm not sure if there are any limits somewhere in your workflow that could silently truncate the attribute contents. Even if that is not the case, my experience is that having so much data in each feature can seriously slow down the workspace.

Then there is also the Sharepoint API to contend with.

My recommendation is to try and isolate more precisely where the problem occurs, and ideally also look if it's possible to streamline the process somewhat, e.g. hold smaller amounts of data in attributes.