Structured Data Matching Improves Your Search Engine

How Structured Data Matching Adds New Features and Improves Usability of Full-Text Search Engine

Structured Data Search Improves Full-Text Search Engines: find out how to develop
and design the ultimate full-text and database search application.

Before choosing the best search engine technology it is important to consider how core search systems are typically classified in two fundamentally different categories.  One category of search engine software was developed for searching full-text documents (web pages on the web and documents on enterprise networks).  The other category of search engine technology specializes in indexing and searching structured data (tables, relational databases, directories, catalogs). 

This white paper provides an overview of the principal differences between these two approaches and how to combine the two for optimal performance.  There are many variables in the mind of a search engine buyer these days:  usability, speed, hardware costs, configurability, leveraging of the latest search enhancements (machine learning, facets, etc.), handling of taxonomies, handling of spelling and synonyms, etc.  Application requirements for search engines rarely take this structured/unstructured data dimension in account fully and cannot be satisfied because the underlying search technology sought by users isn’t adequate to the data they want to access and the way they want it accessed.

In this white paper, Exorbyte presents how it has discovered radically more efficient ways of matching structured data which could never be attempted with full-text unstructured documents.  These approaches offer ways to locate data in any language, in very large data sets, at very high speed, and with a level of control that goes beyond most search software's capabilities:  down to the actual strings of characters contained in each key value pairs of each data record in enormous databases, not only locating exact matches but measuring how close all other near-matches are.  The Exorbyte architecture allows the use of Levenshtein edit distance and other advanced string matching algorithms, phonetics, and Semantics combined with advanced scripted rules.

Finally, this paper explores how such software (Exorbyte MatchMaker) can be combined with existing search infrastructures (including traditional full-text search engines) to transcend the widely accepted limits of search engines today.

Table of Contents

About the Authors

The Exorbyte Inc. team has accumulated an impressive amount of experience with structured data search, search interfaces, and search engine technology.  The author of this particular paper, Daniel Nicollet, is a 15 year veteran of the search software industry (Netcenter, Infoseek, and RealNames).  He has synthesized the incremental search knowledge accumulated by Exorbyte’s developers, interface designers, and business-focused project managers.

About Exorbyte

Exorbyte publishes high-performance search software for all databases and structured data formats.  Exorbyte software offers a large array of phonetic and algorithmic fuzzy search methods while returning results under 10 ms across many millions of records. Exorbyte GmbH was founded in 2000 is headquartered in Konstanz, Germany.  Exorbyte Inc. is a wholly owned subsidiary located in Portland, Oregon.  Exorbyte also has offices in the UK.