Variable-depth trie index optimization: theory and experimental results

Abstract

We develop an efficient approach to Trie index optimization. A Trie is a data structure used to index a file having a set of attributes as record identifiers. In the proposed methodology, a file is horizontally partitioned into subsets of records using a Trie index whose depth of indexing is allowed to vary. The retrieval of a record from the file proceeds by "stepping through" the index to identify a subset of records in the file in which a binary search is performed. This paper develops a taxonomy of optimization problems underlying variable-depth Trie index construction. All these problems are solvable in polynomial time, and their characteristics are studied. Exact algorithms and heuristics for their solution are presented. The algorithms are employed in CRES-an expert system for editing written narrative material, developed for the Department of the Navy. CRES uses several large-tovery-large dictionary files for which Trie indexes are constructed using these algorithms. Computational experience with CRES shows that search and retrieval using variable-depth Trie indexes can be as much as six times faster than pure binary search. The space requirements of the Tries are reasonable. The results show that the variable-depth Tries constructed according to the proposed algorithms are viable and efficient for indexing large-to-very-large files by attributes in practical applications.

Burst tries

“…Some of the proposals discussed above can be re-interpreted in the light of our proposals, as our work and these proposals share the idea of-in our terminology-storing suffixes of strings in containers [40,46]. Another related proposal is the b-trie [49], where b is a block size (such as a disk block), allowing the trie to be used as a disk access structure.…”

“…A burst trie can be viewed as a generalisation of other proposed variants of trie [49,40,46], as discussed later. Figure 1 shows an example of a burst trie storing ten records whose keys are "came", "car", "cat", "cave", "cy", "cyan", "we", "went", "were", and "west" respectively.…”

“…Other approaches to trie compaction are, likewise, only applicable to static tries [3,37,40]. Al-Suwaiyel and Horowitz propose conserving memory by overlaying nodes of a trie so that empty pointers in one node occupy the same location as full pointers in another [3].…”

“…Ramesh et al [40] investigated the problem of finding a good compromise between memory usage and low access costs in a trie. The idea is to collapse subtrees of trie nodes completely and store the (suffix) strings represented by the subtree in a BST that is used as a leaf node of the trie.…”

An efficient modularized database structure for a high‐resolution column‐gridded Mars global terrain database

“…The Terrain relation would not be indexed in storage on the primary key, however; instead a secondary key index for the composite attribute (lat long) would be used, since data would almost always be retrieved on the basis of integer latitude longitude values, that is, for a 1-degree curved rectangle. Given the size of Terrain, some index optimization may be required in practice [41].…”