Moving from bounded to unbounded is fine and is supported, but what I meant is changing the max_size - that would cause corruption in your data. If all you did was change the size to Unbounded then this, in theory, shouldn’t cause any issues.
I’d echo your statement that the schema can be made simpler and more efficient. It’s not exactly clear to me what the key of FILE_CONTENTS is. Is that the file name? Also, are there many file IDs for a single file?