The Index Table

The index table provides a means of locating and searching documents stored in the source table. The index table is maintained by the Full-Text Search engine and has an id column that maps to the IDENTITY column of the corresponding source table. The IDENTITY value from the row in the source table is stored with the data in the Verity collections, which allows the source and index tables to be joined. Although the index table is stored and maintained by the Full-Text Search engine, it functions as a proxy table to Adaptive Server through the Component Integration Services feature.

The index table contains special columns, called pseudo columns, that are used by the Full-Text Search engine to determine the parameters of the search and the location of the text data in the source table. Pseudo columns have no associated physical storage—the values of a pseudo column are valid only for the duration of the query and are removed immediately after the query finishes running.

For example, when you use the score pseudo column in a query, to rank each document according to how well the document matches the query, you may have to use a score of 15 to find references to the phrase “small Saint Bernards” in the text database. This phrase does not occur very often, and a low score value broadens the search to include documents that have a small number of occurrences of the search criteria. However, if you are searching for a phrase that is common, like “large Saint Bernards,” you could use a score of 90, which would limit the search to those documents that have many occurrences of the search criteria.

You use the score column and the other pseudo columns, id, index_any, sort_by, summary, and max_docs, to specify the parameters to include in your search. For a description of the pseudo columns, see “Pseudo Columns in the Index Table”.