AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Benefits of Parallelism in Query Processing
The overhead of sending that data over the network is likely to dominate the time. If you're doing a heavy query and then not returning most of it, it's bad. You return just one row, like average aggregate, but inside it, you need to process a lot of buffers. And this explains why it's so high. I was a bit surprised when you said, like, I get impression you don't turn off parallelism for some of your queries. But would you, if you had the choice, do you think? It depends.