When loading (inserting) data into a table with COPY I have read in the documentation that rows are appended to the end of the table instead of being added to existing table pages, so I’m wondering about memory utilization. Our application uses a number of COPY statements in parallel, so COPY performance is key. Does COPY use the shared buffer pool, or does it allocate its own block of memory to bulk-load the data into?
Would tuning shared_buffers in postgresql.conf have an effect on COPY performance, or is the buffer-pool bypassed altogether on a load?
Thanks,
Keaton