BLOB streaming at SQL level rather than driver level about relstorage HOT 2 CLOSED

Thanks for the feedback, and thanks for the driver!

I wonder if it's an option to do streaming at the SQL level (using the large objection functions) rather than db driver level?

Perhaps a better description of that column would be "Native Streaming Blobs", because, in fact, using the large object functions via SQL is exactly what RelStorage does on pg8000 (or when using psycopg2 in async mode). But this has to divide an upload or download into chunks and perform iterative copying at the Python level (instead of natively in libpq at speed), and IIRC there are differences at the protocol level too. Ultimately it's closer to what MySQL does ("No, emulated via chunking.") than what psycopg2 does (even though the _lobject.py file provides the same interface on pg8000).

from relstorage.

Cool, I'll close the issue, thanks for the quick response 👍

from relstorage.