Interesting they touched on a couple of pain points I’ve had with spark execution, not quite the same thing but I’ve wondered why some of the way it execute isn

Interesting they touched on a couple of pain points I’ve had with spark execution, not quite the same thing but I’ve wondered why some of the way it execute isn’t smarter. Bloom filters for cheap way to filter big joins, using single scans to do multiple jobs in the plan.. I’ll have to have a look.