You can choose to generate dimensions in your Endeca records automatically, basing the dimensions on properties in your source database records.
The following two-dimensional table gives an example of what your source data might look like:
In this source data table, the following relationships exist:
Each row corresponds to a source record. Each source record can be transformed into an Endeca record.
Each column represents a property. Each property can be mapped to a dimension, an Endeca property, or both.
The column entries represent the property values that are assigned to each record. These values map to either dimension values (if the source property is mapped to a dimension), or Endeca property values (if the source property is mapped to an Endeca property).
With automatic dimension generation, source properties become dimensions and source property values become dimension values. If we mapped the Wine Type and Country source properties to dimensions, and then automatically generated those dimensions, the logical representation of the generated dimensions would look like this:
The Wine Type and Country source properties become the Wine Type and Country dimensions. Each source property value becomes a dimension value within its respective dimension: Red, White, and Sparkling for the Wine Type dimension and USA, France, and Chile for the Country dimension. The individual records are organized beneath their respective dimension values accordingly.
With automatically-generated dimensions, it is important to remember that changes to your source data’s properties and property values directly influence those dimensions and their dimension values.
Note
The Information Transformation Layer (ITL) enables you to map source record properties to Endeca dimensions and properties in any way that you require. For example, you can map only some of your source properties to Endeca properties and dimensions, and you can create Endeca properties and dimensions that do not exist in your source data. You can also configure ITL to resolve property name conflicts – for example, you can configure ITL to create synonyms for dimension values that would be identical to their siblings, which is not allowed.