|< < 40 > >|

Join optimization and dynamic programming

Dynamic programming

  • Keep all of the best 1-table plans, one plan for each table.

  • Extend each 1-table plan. For each pair of tables, keep the best plan.

  • Extend each 2-table plan. For each trio of tables, keep the best plan.

  • Compared to the greedy approach, more options are considered at each iteration.

|< < 40 > >|