There are three types of text index that may be created with Texis.

This is the most common index type. It contains a full inversion and complete positional and frequency information of the input data specified by the "set indexexp" indexing regular expressions. This index should probably not be used when the output will be ordered by some other non-relevance-rank factor.

This index is most commonly accessed by using the "LIKEP" SQL clause. It may also be used with LIKE, LIKE3, or LIKER, but since these clauses do not make use of the positional information would generally be a waste of time and disk space to use a full inversion with these clauses. A possible exception to this rule would be if there were multiple LIKE clauses in the query, and one of them is a LIKEP but the rest are either LIKE, LIKE3, or LIKER.

This index type is a low-resolution inversion index that is generally used in conjunction with the LIKE, LIKE3, or LIKER clauses. The index contains frequency and rough positional information. The key advantage to it is that it is very small and very fast at performing look-ups. A typical use for this index type is an "Author, or Title" keyword search where the output is being ordered by a factor other than relevance ranking. Other uses for this index type include multiple LIKE clause statements where go/no-go keyword matching is to be applied to one or more of the fields.

This index type is rarely used but very powerful. The only reason to create this index is when you are using the "LIKEIN" clause. It contains all the same information as a METAMORPH index plus additional proprietary data to support the LIKEIN query type. The index may be also be used as a normal index for querying with the LIKEP, LIKER, LIKE3, and LIKE clauses.