About facets

In your Mashup UI applications you typically use facets through the Refinements panel. By refining your query on a specific facet, you narrow the number of hits displayed in the search results.

Facets are also displayed in charts and table widgets, where you can refine the view by selecting values. For example, in a chart displaying a document count by month, you can focus on the values of a specific month.

This page discusses:

See Also
Create Facets
Numerical Range facets
Date Facets
Multidimension Facets
Geographic Facets
Create Value Facets for Nonhierarchical Metas
Create Aggregations for Facets
Exclusive vs. Disjunctive Refinements

Index-Time and Search-Time Facets

Exalead CloudView has 2 notions of facets:

  • Index-time facets, stored as hierarchical categories in the category field of the index. These index-time facets are often referred to as category facets. They are typically used for alphanumeric metas. After creating index-time facets, you must reindex to store the resulting categories in the category field. In turn, these hierarchal categories display as facets in your search application.

  • Search-time facets, also known as virtual facets. They are based on index fields stored in RAM, and are typically used for dates and numbers. A big advantage of search-time facets is they are available as soon as you create them; you do not need to reindex.

Which Facet Type Should I Use?

The following table explains which type of facet to create, depending on what you want to display in your application. In each section is a cross-reference for learning more about that particular facet type.

Table 1. Facet Types

To display the following

Use this facet type

Dates

Choose from the following Date facets:

  • Date: provides fine-grain control on date sorting.
  • Dynamic dates: Automatically adjusts date granularity, depending on the maximum number of categories specified, and the date range of the search results.

Multiple facets, typically to display in tables, charts, or pivot tables

Choose from the following Multidimension facets:

  • Multi-dimension: Based on N single-dimension facets, and generates a grid, with one facet per dimension. Each cell contains the number of documents and aggregation values for each of the N single-dimension facets. Displays only one level at time, which means better facet performance compared to Hierarchical2D.
  • Hierarchical2D: Based on two single-dimension facets for each axis. Displays all levels in the hierarchy at once. Use for multiple timelines.

Ranged values, such as prices ($0-100, $101-200...)

Numerical ranges are virtual numerical facets used to organize search results in ranges. Specify how to create these ranges:

  • automatically: to specify the number of ranges only, and leave Exalead CloudView define ranges based on search results automatically.
  • that are the same size: to define a fixed size for all ranges. For example, a range of 100 creates the ranges 0-100, 101-200.
  • manually: to define everything, that is, the number of ranges, their max and min values, and the range title to display in the Refinements panel.

Geographic data, either rasters or vectors

Choose from the following Geographic facets:

  • Auto-tile geographic: Use for rasters. It creates geographic facets based on a bounding box. Must be based on a XY or GPS point Data Model property.
  • Geographic: Use for vectors. It creates geographic facets based on disks or polygons. Must be based on a XY or GPS point Data Model property.

Alphanumeric metas with hierarchical values

Category: Use this type of facet if you want to display hierarchical values, or want to define meta mappings to a category manually, instead of using Data Model properties.

This type of facet resides in the index, under the category field in the format. It is the only facet type that requires reindexing to be available in your application.

Alphanumeric metas with nonhierarchical values, such as a list of countries, or file formats

Value: Use this instead of a Category facet. It is saved under the value field, which is optimized for faster faceting.

Only use for nonhierarchical metas. Create this facet on a value type index field.

Concept of Category Facet

All facet types output as a tree made up of facet value objects, called "Category".

In addition, in the data model, when a property is defined as a category facet, such a facet is always created from the values of the properties. For more information about properties, see Using Properties to Configure Document Metas.

Even facets created for numerical data model properties, are always category facets. This means that each value of the property creates a category.