Types of Text Query

There are two primary types of text query that can be performed by Texis. The first form is used to find those records which match the query, and is used when you only care if the record does or does not match. The second form is used when you want to rank the results and produce the best answers first.

The first type of query is done with LIKE, or LIKE3 to avoid post processing if the index can not fully resolve the query. The ranking queries are done with LIKEP or LIKER. LIKER is a faster, and less precise ranking figure than the one returned by LIKEP. The ranking takes into account how many of the query terms are present, as well as their weight, and for LIKEP, how close together the words are, and where in the document they occur. Most queries will use LIKE of LIKEP, with LIKE3 and LIKER used in special circumstances when you want to avoid the post-processing that would fully resolve the query.

There are also two forms of Metamorph index, the Metamorph inverted index and the Metamorph index. The inverted form contains additional information which allows phrases to be resolved and LIKEP rankings to be calculated entirely using the index. This improved functionality comes at a cost in terms of space.

If your queries are single word LIKE queries then the Metamorph index has all the information needed, so the additional overhead of the inverted index is not needed.