This section discusses the logical business object or document attributes that are made searchable to the crawler. Selecting attributes that appear in result items is discussed in the Result Items section of this guideline.
These business object and document attributes must be searchable:
- ID, Short Name, or equivalent
- Attributes providing a descriptive name, such as Name, Long Name, Title, or Description
Additional attributes should be selected based on user needs. All attributes about which users are likely to ask questions or on which users want to filter must be searchable.
Prioritizing or ranking fields enables keyword matches in key attributes, such as Name, to outrank matches to less central attributes, such as Address.
Select the five or six (approximately) most representative and important attributes of your business object or document to which to assign priorities. You must include object identifiers, such as Name and ID. Including Description and Label fields is strongly recommended. Select additional fields for ranking based on business questions that users are likely to ask.
Assign priority values from 1 to 10 (10 is most important) to selected fields based on how likely they are to identify a particular result item to users. Place high priorities (for example, 10) on identifier attributes, such as Name. Fields, such as Long Name and Description, should also receive high priority (for example, a value of 8 or 9). Place lower priorities on fields with values that can include strings that could be associated with different business objects, such as Address. The lowest priority values (1 to 3) should be assigned to attributes that are the least important to users.
Searchable object definitions include keywords. The list of keywords for each searchable object should include:
- Synonyms for the object name
- Synonyms for key object attributes
- Common abbreviations of the object name and attributes
- Synonyms for common domain-specific terms likely to appear as attribute values
- Synonyms for field labels corresponding to the object name or names of its attributes
Global search result items are displayed in a one-column ADF table with horizontal scrolling disabled. Each item is made up of four sections:
Here is an example of the four sections of a result item:
|Sample Result Item
||Time Card : 123456
||TJOHNSON; Submission Date: 01-Apr-2008; Type: Vacation; Amount: 32 hours
||Tags: timecard vacation; Category: Full Time - Regular; Submission Date: 01-Apr-2008; Type: Vacation; Amount: 32 hours; Reason: Not Applicable; Vacation Remaining: 300 hours; Manager: John Smith; Status: Submitted
||Approve; Reject; Return
|Figure 4. Global search results
The main link (default action) is required and is the first (top-leftmost) component of each result item. This is the link to the Oracle Fusion Applications task flow that displays details for the object (for example, purchase order or expense report) represented by the result item. This is the least secure task flow that displays details of the item. Typically, this is a read-only details display. For objects that do not have read-only details, link to an editable object details task flow.
The main link takes this form: <Object Name>: <Object Instance Identifier>
Example: Expense Report: W0020056
The main link can comprise only the object instance identifier if that value provides sufficient information about the result item for users scanning the list of results. Additional attribute values, including additional object instance identifiers (for example, Number plus Name) may be appended to the object instance identifier if they are needed to provide sufficient information about the result item. Examples of object instance identifiers include Number, ID, Person Name, Subject, and Title.
The fixed content (title) starts on the same line as and to the right of the main link, separated from it by white space. The fixed content is made up of a set of attribute-value pairs that appear for every result item that represents the same object type (every result item from the same data source).
Select attributes to include in the fixed content based on these user needs:
- Importance to users: Attributes that users most want to see the values of
- Decisions to act: Attribute values that users use to decide whether to act on the object, such as status, age, amounts, and important dates
Although the fixed content will wrap if its length exceeds the width of the table, it is recommended that you limit the fixed content to five or fewer attributes, all of which should fit on one line in US English.
Developers enter Groovy expressions representing the attribute names and values of all searchable attributes in the form of a single, semicolon-delimited string. The OSES Keyword in Context feature formats the keyword (in bold style) and retrieves 200 bytes of surrounding text to present in the results display. This text, generated by OSES Keyword in Context, forms the variable content of each result item. For example, a search on the keyword vacation would return a string similar to the following:
Employee: Susan Jones; Employee Category: Full Time - Regular; Submission Date: 01-Apr-2008; Type: Vacation; Amount: 32 hours; Reason: Not Applicable; Vacation Remaining: 300 hours; Manager: Philip Smith; Status: Submitted
Tags are searchable and appear at the beginning of the variable content before conventional searchable object attributes. Individual tags are delimited by spaces. Tags appear in boldface if they match a keyword and are included in the 200 bytes retrieved by OSES Keyword in Context. Even though tags are appended to the beginning of a searchable document, it is possible for a tag to match a keyword, but not appear in the result item due to variations in the output of the algorithm that generates the text.
Attributes and Values
Here are the display rules for attributes and values in the Fixed Content and Variable Content sections:
- Use headline capitalization for attribute names and values.
- Include values alone only if their meaning is obvious to users. This is often the case for titles, which are long and descriptive. When the values of multiple attributes have the same data type or are otherwise confusing, their attributes must appear as labels (for example, Created: 01-JAN-00 - Last Updated: 02-FEB-02). If the meaning of the value is vague (that is, not instantly recognizable by users), use the attribute name as a label.
- Separate attributes and values by a colon and a space.
- Separate attribute-value pairs by a semicolon and a space.
The default action for any given result item is what happens when users click the main link. Other actions are optional and appear as links in a row below the variable content. The link text should use the same command name or text that is used on transactional pages that support that action. Add links only for actions that users can decide to take (or not) based on the information available in the global search results UI. Examples of actions include Edit, Submit, Approve, and Forward.
Facets are orthogonal categories or attributes used to filter a body of data. For example, shoes have heel height, color, and size attributes. Most shoppers are interested only in shoes of a certain size or color. A vendor may create a Color facet with values such as black, brown, and red, and then enable users to filter the list of shoes to show only those having the desired value of Color, say black.
Oracle Fusion Search supports two types of facets:
- An attribute of a document or logical business object can be used as a facet. In this case, the facet is an alternate representation of a list of values. Selecting a facet value for one data source removes all other data sources from the results set and displays only results from the selected data source that have the selected value of the selected attribute. For example, selecting the value Escalated on the Service Request Status facet removes all data sources except service requests from the results. Selecting this value also filters the list of matching service requests to show only those with the status of escalated.
- A facet can be a metadata object that is mapped to one or more attributes of a data source. For example, you can map a hierarchical Time facet with the levels Year, Month, and Day to a Due Date attribute. You can map a hierarchical Geography facet to Country, State or Province, and City attributes.
For each searchable object that you own, nominate at least two attributes to use as facets. Users' needs should determine which attributes are mapped to facets. These attributes must be categorical and should have a small number of values (fewer than 20 values is recommended).
To use Date or Number attributes having continuous values as facets, you must create discrete groups of values to expose as facet values. For example, you can group dates into months, quarters, and years. You can group values of numeric attributes, such as annual revenue, into ranges of values, such as Amount less than 100 - 500, 500 - 1000, as so on. Provide names for the ranges, such as Small, Medium, Large, if doing so is beneficial to users.
Subcategory (searchable object) names appear as facet parents in the filter tree. You should express the names in plural form.
Facet (searchable attribute) names should be concise and descriptive and should not repeat the subcategory name (for more information, see the Labeling Common Attributes Guidelines). Similarly, facet values should be concise and descriptive, but not overly long.
The global search results page includes a faceted browsing UI. Facets are organized hierarchically in a tree by category and subcategory. A subcategory corresponds to a data source in SES. Each facet is owned by a data source or subcategory. Facets are referred to as filters in the UI.
When the global search results UI is displayed, the first category is expanded one level by default. The remaining categories appear collapsed: