There may be many parts of a database that need to be combined in the search. For example, Yellow Pages listings have business names, categories, lists of services and products, locations and many other pieces of data. A user may want to search for a business name only, or category only or category and location or any other such combination. In addition, the user wants to search for particular words within the fields, not necessarily the field data exactly as it exists.

One feature is the ability to see extracted contexts. For example, when searching an encyclopedia, there may be hundreds of references for the user's search. The user does not want to have to sift through volumes of text looking for relevant items. Seeing the context of surrounding words helps to identify articles of interest. The user should be able to easily view the entire article or to control the the number of surrounding words in the extract.

The user may want to restrict searches to finding values that are near to each other. For example, a search for Red Bird in an encyclopedia may have thousands of matches. The user might want Red Bird to appear together or to appear within a certain number of words. Also, the user may know that he or she is interested in articles that contain some words but not others.

The user may not know the exact spelling or a search item. The search software should handle wildcards, alternate spellings and Soundex searches. Also, the data may be in alphabets other than English. European languages with accented characters should be searched seamlessly. Non-Latin alphabets should also be supported.

Web search engines typically do not support all of the above features or do not do so very well. A search may return many documents but it may be hard to determine the relevance of each one. A quick test of a popular search engine returned hits that did not include any of the requested values in its displayed list. The user must actually go to the web site and manually look through it for the wanted information. This can be very time consuming and frustrating.

Web search engines search Internet pages. If an organization wants to provide search ability to its own data, it needs to find a product that is viable for such a service.

There are few alternatives in the marketplace that provide such technology that is flexible and easily integrated into existing or legacy data.

Traditional relational databases do not support text search features very easily. The Like clause of the SQL Select statement may not make use of any indexing and thus can be very slow for large databases. Some database systems support text indexing but the searching abilities are very limited.

TxSearch is designed to overcome these limitations and provide flexibility and easy integration with organizational data.

• Library Catalogs
• Reference: Encyclopedias, Dictionaries, and Thesauri
• Sales Catalogs
• Classified: For Sale, Employment, Real Estate
• News, Newspapers, News Groups
• White/Yellow Pages
• People Search, Geneology
• Document repositories, upload and download.

• Wildcard, Soundex, root searching
• Synonyms, alternate spellings
• 8bit Latin alphabets
• Proximity searching
• Boolean searching
• Context extraction
• Hierarchical data