Reducing Order Enforcement Cost in Complex Query Plans

Abstract

We also executed our plans on a widely used commercial database system, and on PostgreSQL, and found that actual execution times for our plans were significantly better than for plans generated by those systems in several cases.

The common prefix problem on trees

“…, P |V | , so as to maximize uv∈E |P u ∧ P v |. This problem arises in database query optimization [1]. The corresponding decision problem is known to be NP-Complete [1,2] by a reduction from the SUM-CUT problem.…”

“…This problem arises in database query optimization [1]. The corresponding decision problem is known to be NP-Complete [1,2] by a reduction from the SUM-CUT problem. It can be solved in polynomial time if the tree is a path, and a 1/2-factor approximation is known for binary trees [1,2] (see Section 2).…”

“…The corresponding decision problem is known to be NP-Complete [1,2] by a reduction from the SUM-CUT problem. It can be solved in polynomial time if the tree is a path, and a 1/2-factor approximation is known for binary trees [1,2] (see Section 2). In this paper we give a (1 − o(1)) factor algorithm for this problem when the underlying graph is a tree with the vertex degrees bounded by a constant.…”

“…Using the same analysis as above, we get a (1 − ) factor algorithm that runs in O(n2 2 1 +1 ) steps (besides the operation on the labels). In particular, when = 1/2, we get the 1/2 factor algorithm of [1,2].…”

“…However solving this recursion may involve steps exponential in the number of nodes. For example, on a complete binary tree of size n. The recursion-(1) can be implemented as a dynamic program for trees which have a polynomially bounded number of subtrees, for example, paths (see [1,2]). We observe that binary trees of height log log n also have this property.…”