Extremely large indexes require more than one physical partition. If more than one partition exists, each partition is queried individually and the results merged for presentation to the end-user. The mergeSettings attribute works together with responseNumberSettings to control the number of results returned from a merged result set.

The responseNumberSettings attribute determines how many results are returned from each partition. The mergeSettings attribute determines how many of those total results are returned to the end user. If you use mergeSettings, be sure to set it to a number higher than the individual partition results set in responseNumberSettings. For example, if you return 50 results from each of four partitions, you may use mergeSettings to trim the combined result list to the top 100 results, but not the top 20 results. It is more efficient to trim the responseNumberSettings to begin with than to do so after merging.

The mergeSettings attribute takes the same options as responseNumberSettings. If mergeSettings is not set, responseNumberSettings is used instead. The syntax for mergeSettings is:

mergeSettings=doc50,prop50,…

Grouping Type

Description

Group-by-document

Groups the raw search results by document, returning up to some maximum number of groups of a certain size, as defined by these parameters, with the default values shown:

doc10,perDoc3,perSol1,

- doc–Maximum number of document result groups to return

- perDoc–Maximum size of a group from an unstructured index item

- perSol–Maximum size of a group from a structured index item.

Note: An additional mode, docrank, is the same as document, but it also uses the relevancy of the document instead of the relevancy of the statement to rank results.

Group-by-property

Groups the raw search results by a metadata property value, returning up to some maximum number of groups of a certain size, as defined by these parameters, with the default values shown:

prop10,perProp3,

The prop parameter is the maximum number of property result groups to return, and the perProp parameter establishes the maximum size of a group.

To group by property, the mode value requires a sortProp attribute with the type, name, and default value for the grouping property. See the sortProp section in this guide.

Result Type Weights

Normally, all statement results receive the same treatment in the relevancy calculation. However, you may want certain statement types to be weighted higher or lower in the search results. For example, two identical statements from two similar documents usually receive near identical relevancy, with minor differences in the context and document weight factors. However, if the statements are from two different text fields (such as role:goal and role:fact), and these fields were weighted differently, then their relevancy could vary greatly. Oracle ATG Web Commerce Search supports these weighting factors with the following parameters:

f*1.0,o*1.0,s*1.0,ROLE:ID*2.0

f*—The weight (or multiplier) of preferred answer statement relevance.

o*—The weight of structured statement results.

s*—The weight of unstructured statement results.

A weight of 1.0 means the original (pure) relevancy is used.

Individual structured types (or fields) can be defined separately, as shown:

role:goal*1.2,role:symptom*1.1,role:fact*0.5

Whole Field Result Text

Normally, the result text is the matching statement text plus some additional context for small sentences. However, for structured content, which contains potentially multi-sentence fields of text, you might want to return the entire text of the field as the result. This behavior is controlled by the following parameter:

wholefield0

The wholefield parameter holds a Boolean value which, if non-zero, means that the entire enveloping field text for the result’s matching statement is returned as the text.

Suppress Similar Statements

When using Oracle ATG Web Commerce Search with Oracle ATG Web Knowledge Manager, you may want to retrieve documents that have unique matching text. Search allows you to suppress documents from the current page if they share a matching text statement. This behavior is controlled by the following parameter:

suppress0

The suppress parameter holds an integer value that determines the number of matching statements to compare for suppression. For example, if the value is 2, then the top-2 matching statements for each document result are compared. If two documents share a matching statement, the first one on the page is returned, the other is suppressed. Note that the comparison is case-sensitive and ignores whitespace and punctuation.

Note that the total number of results that is reported from the engine does not take into account this suppression.