Source property mappings dictate which dimension values are tagged to each record and which property information is available for record search, sort, and display.
Source properties can be mapped in three different ways. They can be:
Note the following requirements when you map properties:
You use a property mapper component to establish source property mappings. Typically, the property mapper is placed in the pipeline after the Perl manipulator that is used to clean and prepare source properties (if one exists). You should use a single property mapper to map all of your source properties to both Endeca properties or dimensions.
Note
Before you can map a source property to an Endeca property or dimension, you must create the Endeca property or dimension.
Understanding property mappers is critical to building an Endeca implementation. Oracle recommends that you read the chapter "Mapping Source Properties" in the Endeca Forge Guide to gain a solid grounding in this critical concept.
You should use a single property mapper to map all of your source properties to both Endeca properties or dimensions.
Note
There are rare cases where multiple property mappers may be used. However, Oracle strongly recommends that you use only one property mapper in any given pipeline.
When you specify a source property and a target Endeca property or dimension to map to, you are creating an explicit mapping.
In general, this is the type of mapping Oracle recommends you use. However, Endeca also offers some advanced techniques that allow you to automate the mapping process. These techniques are intended to facilitate the process of building prototypes and should not be used for building production-ready implementations.
See the related topics for detailed explicit mapping procedures.
When Forge maps a source property value to a dimension value, the dimension value it uses can either be explicitly defined in the dimension hierarchy or automatically generated by Forge. You set the type of dimension value handling, on a per mapping basis, using the Match Mode in the Dimension Mapping dialog box.
Note
Match modes only apply to dimension values. They are not used when mapping source properties to Endeca properties.
There are three match modes you can choose from:
Normal maps only those source property values that have a matching dimension value explicitly defined in the dimension hierarchy. Forge assigns the IDs for any matching dimension values to the Endeca records. Any source property values that do not have matching dimension values in the dimension hierarchy are ignored.
In order for a source property value to match a dimension value, the dimension value's definition must contain a synonym that:
Note
There are two types of advanced dimension values, range and sift. These dimension value types have different matching criteria than standard dimension values. See About range dimension values and About sift dimensions, respectively.
Must Match behaves identically to Normal, with the exception that Must Match issues a warning for any source property values that do not have matching dimension values.
Auto Generate specifies that Forge automatically generates a dimension value name and ID for any source property value that does not have a matching dimension value in the dimension hierarchy. Forge uses these automatically-generated names and IDs to tag the Endeca records the same as it would explicitly-defined dimension values.
Auto Generate dramatically reduces the amount of editing you have to do to the dimension hierarchy. However, auto-generated dimensions are always flat. Auto-generated names and IDs are persisted in a file that you specify as part of a dimension server component.
When you choose which method to use for generating your dimension values, there are two rules of thumb to keep in mind:
If you manually define dimension values in the dimension hierarchy, the Normal, Must Match, and Auto Generate features behave identically with respect to those dimension values.
Any source property value that does not have a matching dimension value specified in the dimension hierarchy will not be mapped unless you have set the dimension to Auto Generate in the pipeline.
The following illustration shows a simple dimension mapping example that uses a combination of generation methods. The sections after the illustration describe the mapping behavior in the example.
The Red and White property values have matching Red and White dimension values specified in the dimension hierarchy. These property values are mapped to the Red and White dimension values IDs, respectively. Bottles B and C are tagged with the Red dimension value ID, and Bottle A is tagged with the White dimension value ID.
The Sparkling property value does not have a matching dimension value in the dimension hierarchy. The Wine Type dimension is set to Must Match, so this property is ignored and a warning is issued. As a result, Bottle D does not get tagged with a dimension value ID from the Wine Type dimension.
There are no dimension values explicitly defined in the dimension hierarchy for the Country dimension. However, this dimension is set to Auto Generate, so all three of the Country property values (USA, France, and Chile) are mapped to automatically-generated dimension value IDs.
Bottle A is tagged with the auto-generated ID for the USA dimension value. Bottles B and D are tagged with the auto-generated ID for the France dimension value. Bottle C is tagged with the auto-generated ID for the Chile dimension value.
The Crisp property value has a matching dimension value specified in the dimension hierarchy, so the Crisp property value is mapped to the Crisp dimension value. Bottle A is tagged with the Crisp dimension value ID.
The other three property values (Elegant, Full, and Fresh) do not have matching dimension values in the dimension hierarchy but, because the Body dimension is set to Auto Generate, these three property values are mapped to automatically-generated dimension values IDs.
Bottle B is tagged with the auto-generated ID for the Elegant dimension value. Bottle C is tagged with the auto-generated ID for the Full dimension value. Bottle D is tagged with the auto-generated ID for the Fresh dimension value.
Regardless of how they were generated, all of the dimension value IDs are included in the finished data that Forge produces for indexing.
You can assign more than one mapping to a source property-for example, you can map a source property to both a dimension and an Endeca property.
A typical source property that you may want to map to both a dimension and an Endeca property is Price. You can map the Price source property in the following ways:
Conversely, you can assign more than one source property to a single dimension or Endeca property. For example, if you have multiple source properties that are equivalent, most likely they should all be mapped to the same dimension or Endeca property. Flavor and Color are example properties that might require this behavior