A business depends upon finding and exploiting information, most of which is stored digitally in the form of text. However, the number of documents and the volume of text data continues to grow apace, making it increasingly difficult to find vital facts and figures. When conventional tools are employed, the search for documents is often slow and haphazard.
However, according to Joe Rios, Software AG business manager at SPL, there is an answer to this problem. "Adabas from Software AG is a powerful, proven database designed to accommodate a wide range of data types in the enterprise environment, including text. Adabas is available with an optional extension: Adabas Text Retrieval.
"This add-on module combines an advanced indexing facility with a sophisticated query processor to make document searches in Adabas faster and more accurate, and allows users to define their information requirements with greater precision," he comments.
A full-text facility allows users to index documents of practically any kind, both formatted and unformatted. Says Rios: "It is possible to define terms to be excluded from the index to reduce system load. If the number of terms required is inherently limited, you can also create a positive inclusion list, greatly increasing performance. Indices can therefore be tailored to the client application and end user needs."
The retrieval process is fast and flexible - because it is based on content (indexed data), not laborious searches through multiple original documents. It is possible to search for particular words or strings, combinations or patterns, for alphanumeric attributes such as 'date document created', or for practically any combination. It is also possible to perform a more refined search on an existing query result, allowing the user to bring large numbers of documents down to manageable levels.
Adabas Text Retrieval enables searches to be defined in the following ways:
Search for a given word within a specified distance or a second word, irrespective of order; Search for a given word within a specified distance of a second word, in a defined order; Search for a given word within same or specified number of sentences or paragraphs of a second word; Search for words of the same meaning but with different spellings; Search for a document containing a given word or a predefined synonym, more generic term or more specific term; Search using logical operators (AND, OR, NOT); Search using document attributes (alphanumeric data, such as author, dates) or combinations of attributes and actual text (eg find documents created by 'x' on day 'y' containing word 'z'); and Search for truncated words or letter combinations.
"This allows users to leverage their knowledge of context and relationships between search terms, as well as their understanding of text structure," says Rios. "It also makes it possible to perform 'fuzzy; searches based on very general search criteria and then refine them.
"In other words, Adabas Text Retrieval makes the process of finding information stored in documents much easier, more precise and faster," he adds.
Adabas Text Retrieval manages index information (meta data) and not the content of text files themselves. As a result, it can be combined with diverse document management or word processing applications, and can be employed to index many types of data and text stored on a variety of media (eg files stored in Adabas, sequential files, even on CD-ROM).
The call-level interface for Adabas Text Retrieval can be embedded in any third-generation language, such as C, Cobol or PL/1. It can also be embedded in Natural, Software AG's platform for business applications. Natural Document Management includes Adabas Text Retrieval, guaranteeing a seamlessly integrated solution from the word go.
Indexed data is stored in Adabas - ensuring data independence, automatic backup and recovery, high performance and high capacity. "These advantages can, of course, also be extended to the document files themselves, by storing them too, in the tried and trusted Adabas environment. As text data is stored independently from the applications, it is available on a variety of solutions - a highly important consideration in today's complex, distributed information technology landscapes," says Rios.
Adabas Text Retrieval is already in productive use around the world. Customers include the New South Wales Police Department in Australia, the ZDF television corporation and Boehringer Pharmaceuticals in Germany, Commercial Union Assurance in the UK and Telef'onica de Espa~na in Spain.
Share