H Stopword Transformations

H.1 Understanding Stopword Transformations

When you use a stopword or stopword-only phrase as an operand for a query operator, Oracle Text rewrites the expression to eliminate the stopword or stopword-only phrase and then executes the query.

The following section describes the stopword rewrites or transformations for each operator. In all tables, the Stopword Expression column describes the query expression or component of a query expression, while the right-hand column describes the way Oracle Text rewrites the query.

The token stopword stands for a single stopword or a stopword-only phrase.

The token non_stopword stands for either a single non-stopword, a phrase of all non-stopwords, or a phrase of non-stopwords and stopwords.

The token no_lex stands for a single character or a string of characters that is neither a stopword nor a word that is indexed. For example, the + character by itself is an example of a no_lex token.

When the Stopword Expression column completely describes the query expression, a rewritten expression of no_token means that no hits are returned when you enter such a query.

When the Stopword Expression column describes a component of a query expression with more than one operator, a rewritten expression of no_token means that a no_token value is passed to the next step of the rewrite.

Transformations that contain a no_token as an operand in the Stopword Expression column describe intermediate transformations in which the no_token is a result of a previous transformation. These intermediate transformations apply when the original query expression has at least one stopword and more than one operator.

For example, consider the following compound query expression:

'(this NOT dog) AND cat'

Assuming that this is the only stopword in this expression, Oracle Text applies the following transformations in the following order:

stopword NOT non-stopword => no_token

no_token AND non_stopword => non_stopword

The resulting expression is:

'cat'