S2024 #13 - Query Optimizer Implementation 1 (CMU Advanced Database Systems)

แชร์
ฝัง

ความคิดเห็น • 1

  • @akurmustafa_
    @akurmustafa_ 6 หลายเดือนก่อน +1

    As always very nice lecture. Got a question though. I think representing joins (inner, cross joins. Might be cumbersome for left-right kind of joins) as a single logical plan operator with possibly more than 2 children is helpful for subsequent optimization stages. By this way, result of the cost-based optimizations can be calculated with less iteration. I also think this approach would be better for filter predicate push-down stage. I wonder whether there is any system having this approach, or a paper discussing it?