Document Management with Oracle Text

June 12, 2003

A few years ago, many projects started using the multimedia capability of Oracle databases. Powerful procedures for handling image, audio, video and the text data in various formats are incorporated in the database kernel code; even the default database installation has multimedia objects installed. Storing, retrieving and searching the text from documents stored inside the Oracle database are the most common usage of multimedia database capabilities. This feature, incorporated into the product was formerly known as ConText, later known as interMedia Text (iMT) and with the database version 9.X, it is known as Oracle Text.

This article covers:

  • Document Index Type
  • Oracle Text Architecture (Classes, Objects, Preferences, Attributes)
  • Oracle Text Installation Check
  • Text Indexes Inventory
  • Synching and Optimising a Text Index
  • Monitoring and Error Logging
  • Maintenance Tips
  • Conclusion

Document Index Type

Oracle Text is an extension to the Oracle database that allows searching specific words in the tables of documents, using standard SQL expressions. Oracle Text is integrated in a number of Oracle products such as Portal, iFS and Applications. Supported document types are text, HTML, DOC, XLS, PPT, PDF and XML documents. txt/HTML data content will be stored in the clob column and other formatted document content in the blob column. The content can "also" stored outside of database via BFILES. For any kind of data content, the text engine is used for indexing and retrieving.

Overview of the Oracle Text product development:

Product Name

Database version

Index Type

ConText 2.x.x

< 8.1.5

interMedia Text

< 8.1.7

Context

Oracle Text for Oracle8i

(V 8.1.x) 8.1.7

Catalog, Context

Oracle Text

9.x

Ctxxpath, Catalog, Context

There are three different index types: Context, Catalog and Ctxxpath. They are used all for document indexing, but each has a different functionality.

Context index is a "domain" index used for fast retrieval of unstructured text.

DML processing on a Context index is deferred. The actual index updates do not take place until an index SYNC is performed.

Catalog (CTXCAT) index is an online, "catalog" index, efficient for searching between small, simple text fields and with queries using some structured criteria, (usually numbers or dates). This index type supports only a basic functionality provided in a Context index. A Catalog index has all the characteristics of the normal database index.

Ctxxpath index is a special index installed during an Oracle Text install. This index uses Oracle Text code and can be created only on sys.xmltype columns. It is used to speed up certain queries using the existsnode method.








The Network for Technology Professionals

Search:

About Internet.com

Legal Notices, Licensing, Permissions, Privacy Policy.
Advertise | Newsletters | E-mail Offers