---

clusterTree

Clusters provide users with a tree structure to navigate the top n results by organizing search results into groups. Documents in the same group share the same common topics or property values. Effective real-time clustering balances clustering quality and clustering time.

Clustering does not change the order of the documents. When users select a cluster, the result view is limited to the documents in that cluster. All operations, such as sorting or next page, are limited to the cluster.

Object Type

Creatable

Object Key

name

Object Key Command Syntax

--NAME=object_name

-n object_name

State Properties

Property	Value
status	ACTIVE INACTIVE

Supported Operations

activate
create
createAll
deactivate
delete
deleteAll
deleteList
export
exportAll
exportList
getAllObjectKeys
getAllStates
getState
getStateList
update
updateAll

Administration GUI Page

Global Settings - Clustering Configuration - Create or Edit Metadata Clustering Tree
Global Settings - Clustering Configuration - Create or Edit Topic Clustering Tree

XML Description

The <search:clusterTrees> element describes topic cluster trees, metadata cluster trees, or both:

<search:clusterTrees>
  <search:topicClusterTree>
    <search:name>
    <search:translations>
    <search:clusteringAttrs>
      <search:clusteringAttr>
        <search:name>

  <search:metadataClusterTree>
    <search:name>
    <search:translations>
    <search:clusteringAttrs>
      <search:clusteringAttr>
        <search:name>
        <search:type>
    <search:tokenized>
    <search:tokenDelimiter>
    <search:hierarchical>
    <search:hierarchyDelimiter>

Element Descriptions

<search:clusterTrees>

Contains one or more <search:topicClusterTree> elements, <search:metadataClusterTree> elements, or both.

<search:topicClusterTree>

Describes a topic cluster tree. It contains these elements:

<search:name>
<search:translations>
<search:clusteringAttrs>

<search:metadataClusterTree>

Describes a metadata cluster tree. It contains these elements:

<search:name>
<search:translations>
<search:clusteringAttr>
<search:tokenized>
<search:tokenDelimiter>
<search:hierarchical>
<search:hierarchyDelimiter>

<search:name>

Contains the unique name of the cluster tree. Required.

<search:translations>

Contains one or more translations of the object name. See "Providing Translations of Object Names".

<search:clusteringAttrs>

Contains one or more <search:clusteringAttr> elements.

<search:clusteringAttr>

Contains a <search:name> element and, for metadata trees, a <search:type> element.

These attributes can be default search attributes, custom search attributes, or Oracle SES internal attributes. Topic tree attributes are String only. For metadata trees, you must specify the data type.

<search:name>

Contains the search attribute used to generate the tree.

<search:type>

Contains the data type of the attribute values. Set to one of these values: STRING, NUMBER, or DATE.

<search:tokenized>

Controls tokenizing of a String attribute value in a metadataClusterTree. Set to true to separate the string into several values where indicated by a delimiter, or set to false to handle the string as a single value.

<search:tokenDelimiter>

Identifies the delimiter used to separate tokens in a String attribute value. Set to a character, such as a comma (,) or a hash mark (#). The default delimiter is whitespace ( ). The token delimiter must be different from the hierarchy delimiter when both are used.

<search:hierarchical>

Controls whether a metadata cluster tree for String attributes has a hierarchical structure. Set to true to generate the tree based on a hierarchy implicit in the attribute values, or set to false to generate the tree without a hierarchy.

<search:hierarchyDelimiter>

Identifies the delimiter used to separate the categories in a hierarchy for a metadata cluster tree. Set to a character, such as a slash (/). The default delimiter is whitespace ( ). The hierarchy delimiter must be different from the token delimiter when both are used. Tokens are parsed before the hierarchy.

The following example shows a comma-delimited tokens, and both tokens have a three-level, slash-delimited hierarchy:

java/j2ee/jdbc, oracle/search/connector

Example

This XML document defines both a topic cluster tree and a metadata cluster tree:

<?xml version="1.0" encoding="UTF-8" ?>
<search:config productVersion="11.1.2.0.0" xmlns:search="http://xmlns.oracle.com/search">
  <search:clusterTrees>
    <search:topicClusterTree>
      <search:name>Topic Tree</search:name>
      <search:translations>
        <search:translation language="es">
          <search:translatedValue>Árbol del Asunto
          </search:translatedValue>
        </search:translation>
      </search:translations>
      <search:clusteringAttrs>
        <search:clusteringAttr>
          <search:name>eqtopphrases</search:name>
        </search:clusteringAttr>
        <search:clusteringAttr>
          <search:name>eqsnippet</search:name>
        </search:clusteringAttr>
      </search:clusteringAttrs>
      </search:topicClusterTree>
      <search:metadataClusterTree>
        <search:name>Metadata Tree</search:name>
        <search:translations>
          <search:translation language="es">
            <search:translatedValue>Árbol de los Meta Datos
            </search:translatedValue>
        </search:translation>
      </search:translations>
      <search:clusteringAttr>
        <search:name>Infosource</search:name>
        <search:type>STRING</search:type>
      </search:clusteringAttr>
      <search:tokenized>true</search:tokenized>
      <search:tokenDelimiter>,</search:tokenDelimiter>
    </search:metadataClusterTree>
  </search:clusterTrees>
</search:config>