6.4 Create a Text Policy

An Oracle Text policy specifies how text content must be interpreted. You can provide a text policy to govern a model, an attribute, or both the model and individual attributes.

If a model-specific policy is present and one or more attributes have their own policies, Oracle Machine Learning for SQL uses the attribute policies for the specified attributes and the model-specific policy for the other attributes.

The CTX_DDL.CREATE_POLICY procedure creates a text policy.

CTX_DDL.CREATE_POLICY(
          policy_name    IN VARCHAR2,
          				filter         IN VARCHAR2 DEFAULT NULL,
          				section_group  IN VARCHAR2 DEFAULT NULL,
          				lexer          IN VARCHAR2 DEFAULT NULL,
          				stoplist       IN VARCHAR2 DEFAULT NULL,
          				wordlist       IN VARCHAR2 DEFAULT NULL);

The parameters of CTX_DDL.CREATE_POLICY are described in the following table.

Table 6-3 CTX_DDL.CREATE_POLICY Procedure Parameters

Parameter Name	Description
policy_name	Name of the new policy object. Oracle Text policies and text indexes share the same namespace.
filter	Specifies how the documents must be converted to plain text for indexing. Examples are: CHARSET_FILTER for character sets and NULL_FILTER for plain text, HTML and XML. For filter values, see "Filter Types" in Oracle Text Reference.
section_group	Identifies sections within the documents. For example, HTML_SECTION_GROUP defines sections in HTML documents. For section_group values, see "Section Group Types" in Oracle Text Reference. Note: You can specify any section group that is supported by CONTEXT indexes.
lexer	Identifies the language that is being indexed. For example, BASIC_LEXER is the lexer for extracting terms from text in languages that use white space delimited words (such as English and most western European languages). For lexer values, see "Lexer Types" in Oracle Text Reference.
stoplist	Specifies words and themes to exclude from term extraction. For example, the word "the" is typically in the stoplist for English language documents. The system-supplied stoplist is used by default. See "Stoplists" in Oracle Text Reference.
wordlist	Specifies how stems and fuzzy queries must be expanded. A stem defines a root form of a word so that different grammatical forms have a single representation. A fuzzy query includes common misspellings in the representation of a word. See "BASIC_WORDLIST" in Oracle Text Reference.