Import and verify metadata
After identifying the metadata that can be used in your model, import only the required items to keep your model manageable. Also, verify the Usage and Aggregate properties on columns.
The following Usage settings are supported:
- Represents a key, code, or date.
- Represents context, such as name, color, geographic information.
- Represents key performance indicators that are typically aggregated in the context of identifiers or attributes.
By default, the Aggregate property for an identifier or attribute is set to Count, and for a measure to Sum or Total. Even though a fact table might hold a numeric attribute or identifier (for example, an order number), it doesn’t mean that the item in the fact table is a measure, and its aggregation should be set accordingly, such as Count or no aggregation. The same logic applies to a dimension where an attribute, such as population or square footage, might not necessarily warrant an aggregation setting of Sum or Total, but rather the setting of no aggregation.
You typically want to ensure that integer values that represent keys, codes, or non-rollup integer attributes in your data have the Usage property set to Attribute or Identifier, and the Aggregate property set to Count. It's not logical for these values to behave as measures, and be totaled or averaged.
For information on setting column properties in data modules, see Object properties.
For information on setting query item properties in Framework Manager, see Query items.
Data modules can denote that columns are from the domain of geographic (city, state, country, and so on) or temporal types (day, month, quarter, year, and so on). This extended metadata can be automatically leveraged in queries. For example, data values that represent geographic data automatically have their Represents property set to Geographic location. With time-based data, this property is automatically set to Time. Cognos® Analytics also analyzes the data and allows for other functionality, such as data visualization recommendations. Verify the Represents property on columns to ensure that it's set as expected.
For Framework Manager models, you can enrich a published package to achieve the same artificial intelligence (AI) functionality as in data modules.
For more information, see Enriching packages.