What's New in Semantic Technologies?

This section describes new and changed semantic technologies features for Oracle Database Release 11.

Release 11.2 Features

The following are new and changed features for Oracle Database 11g Release 2 (11.2).

This release also includes the features that were supplied in the interim patch 7600122 for Release 11.1.0.7.0, which are listed in "Features Added for Release 11.1.0.7 (November, 2008)".

Semantic Indexing for Documents

You can use semantic indexing to enable queries on information extracted from unstructured documents. Documents indexed semantically can be searched using the SEM_CONTAINS operator within a standard SQL query. For information about using semantic indexing, see Chapter 4.

The SEM_RDFCTX PL/SQL package contains subprograms to manage extractor policies and semantic indexes created for documents. This package is documented in Chapter 9.

Virtual Private Database (VPD) and Oracle Label Security (OLS) Support

You can enforce a fine-grained access control mechanism for RDF data by using either the Virtual Private Database (VPD) or Oracle Label Security (OLS) feature of Oracle Database, as explained in Chapter 5.

Workspace Manager Support for RDF Data

You can use Oracle Workspace Manager to version-enable RDF data in the semantic data store, as explained in Chapter 6.

New Procedures to Enable, Downgrade, and Remove Semantic Technologies Support

The procedures for enabling, downgrading, and removing semantic technologies support in the database have been redesigned. These procedures are explained in Appendix A.

Note that you still must perform certain actions and meet prerequisites before you can use any types, synonyms, or PL/SQL packages related to Oracle semantic technologies support. These actions and prerequisites are explained in Section A.1.

Optimized owl:sameAs Inference

You can optimize the performance of owl:sameAs inference, as explained in Section 2.2.8.

Optimizing Entailment Performance Through Compact Structures

You can specify RAW8=T in the options parameter to the SEM_APIS.CREATE_ENTAILMENT procedure, to cause RAW8 datatypes to be used instead of NUMBER in many intermediate tables created during the inference process. This option can improve entailment performance by up to 30% in some cases.

Incremental Inference

You can use incremental inference update entailments (rules indexes) efficiently after triple additions, as explained in Section 2.2.9.

Parallel Inference

You can use parallel inference to improve inference performance by taking advantage of the capabilities of a multi-core or multi-CPU architectures, as explained in Section 2.2.10.

Filter and Union Support in Curly Brace Syntax for SEM_MATCH

You can now specify FILTER or UNION, or both, in addition to OPTIONAL as keywords in the curly brace syntax supported for the SEM_MATCH table function. These keywords are explained in Section 1.6.2.

Simple Knowledge Organization System (SKOS) Support

You can perform inferencing based on a core subset of the Simple Knowledge Organization System (SKOS) data model, as explained in Chapter 3.

Inference Beyond OWLPrime

You can specify additional inference components to cover OWL constructs, such as owl:intersectionOf and owl:unionOf, as explained in Table 7-1, "Inferencing Keywords for inf_components_in Parameter" in the Usage Notes for the SEM_APIS.CREATE_ENTAILMENT procedure in Chapter 7.

Systematized Nomenclature of Medicine - Clinical Terms (SNOMED CT) Support

The current release includes a built-in rulebase that supports the expressiveness of Systematized Nomenclature of Medicine - Clinical Terms (SNOMED CT) terminology. To include this support, specify the SNOMED keyword in the inf_components_in parameter in the call to the SEM_APIS.CREATE_ENTAILMENT procedure .

New SEM_APIS Package Subprograms

The following subprograms have been added to the SEM_APIS PL/SQL package, which is documented in Chapter 7.

Parameters for SEM_PERF.GATHER_STATISTICS

The SEM_PERF.GATHER_STATS procedure now accepts two parameters, both optional: just_on_values_table and parallel. This procedure is documented in Chapter 8.

SEM_APIS.*_RULES_INDEX Subprograms Deprecated

In the SEM_APIS PL/SQL package, the subprograms containing "_RULES_INDEX" in their names are deprecated and removed from this manual. Instead, you should use subprograms containing "_ENTAILMENT" in their names:

Instead of SEM_APIS.ALTER_SEM_INDEX_ON_RULES_INDEX, use SEM_APIS.ALTER_SEM_INDEX_ON_ENTAILMENT.
Instead of SEM_APIS.ANALYZE_RULES_INDEX, use SEM_APIS.ANALYZE_ENTAILMENT.
Instead of SEM_APIS.CREATE_RULES_INDEX, use SEM_APIS.CREATE_ENTAILMENT.
Instead of SEM_APIS.DROP_RULES_INDEX, use SEM_APIS.DROP_ENTAILMENT.
Instead of SEM_APIS.LOOKUP_RULES_INDEX, use SEM_APIS.LOOKUP_ENTAILMENT.

The old "*_RULES_INDEX" formats will continue to work, but you are encouraged to switch to using the "*_ENTAILMENT" subprograms.

Note that several metadata views still include "RULES_INDEX" in their names.

The terms entailment and rules index are synonyms in this manual, although entailment is used predominantly

Release 11.1 Features

The following are new and changed features for Oracle Database 11g Release 1 (11.1).

Features Added for Release 11.1.0.7 (November, 2008)

This section describes features that are included in interim patch 7600122 for Release 11.1.0.7.0, which was made available on Oracle MetaLink in November, 2008.

Support for virtual models (see Section 1.3.8)
Curly brace syntax for SEM_MATCH graph pattern, including support for the OPTIONAL construct (see Section 1.6.2)
Using HINT0 ("hint-zero" ) in a SEM_MATCH query (see Section 1.6)
New columns returned from SEM_MATCH: id, _prefix, _suffix (see Section 1.6)
Ability to create and manage semantic network indexes on models and rules indexes (see Section 1.8)
SEM_APIS.BULK_LOAD_FROM_STAGING_TABLE procedure: New options in the flags parameter join_hint for tasks IZC, MBV, and MBT; parallel=n)
Simplification of staging table definition (fewer columns) and privilege requirements (see Section 1.7.1)
Simpler event tracing during bulk load (see the information about the new RDF$ET_TAB table in Section 1.7.1.1)

Storage Model Enhancements and Migration

The storage model has been enhanced to support OWL inferencing: some internal data structures and indexes have been changed, added, and removed. These changes also result in enhanced performance.

Because of the extent of these changes, if you have semantic data that you used with the previous release, you must upgrade that data to migrate it to the new format before you can use any new features for this release. For more information, see Appendix A, "Enabling, Downgrading, or Removing Semantic Technologies Support".

Support for OWL Inferencing

Support has been added to support storing, validating, and querying Web Ontology Language (OWL)-based ontologies. Support is provided for a subset of the OWL DL language.

To query ontology data, you can use table functions and operators that examine semantic relationships, such as SEM_MATCH, SEM_RELATED, and SEM_DISTANCE.

New Bulk Loading Interface for Improved Performance

You can improve performance for bulk loading of semantic data in bulk using a staging table and calling the SEM_APIS.BULK_LOAD_FROM_STAGING_TABLE procedure. For more information, see Section 1.7.1.

Ontology-Assisted Querying of Relational Data

You can go beyond syntactic matching to perform semantic relatedness-based querying of relational data, by associating an ontology with the data and using the new SEM_RELATED operator (and optionally its SEM_DISTANCE ancillary operator). The new SEM_INDEXTYPE index type improves performance for semantic queries.

Required Procedure for Using Semantic Technology Support

Before you can use any types, synonyms, or PL/SQL packages related to Oracle semantic technologies support, you must enable support for semantic technologies. For more information, see Appendix A, "Enabling, Downgrading, or Removing Semantic Technologies Support".