To optimize the user experience in IBM® Cognos® Analytics with Watson components, such as dashboards and explorations, Framework Manager packages must be enriched.
The enrichment process associates the Cognos Analytics data characteristics, such as Time and Geographic location, to query items in the packages. The information from the enrichment process complements the information, such as the data type, column name, or Usage property value, that is derived from the package metadata.
An enriched package includes the data characteristics that are required for the artificial intelligence (AI) based functionality in the product, such as visualization recommendations or intelligently set default values on column properties. For example, to display the relationships diagram in Explore, an enriched package must be used. Otherwise, the relationships diagram isn’t displayed.
- Names of query subjects, query items, and namespaces are changed.
- Data types on query items are changed. For example, number changed to string.
- New query items are added.
- Filters or expressions are changed that significantly alter the values that the query subject would return.
- A deployment archive is imported into a new environment that uses different data from the source used for a previous enrichment.
When a package is republished, existing enriched metadata isn’t removed or refreshed.
Before you begin
To minimize the impact of the enrichment process on the system, consider creating smaller packages that include only a subset of purpose-specific query subjects, and enriching only the smaller packages. For example, a package used by advanced report authors might expose many query subjects where many of the query subjects aren’t relevant when creating dashboards or explorations. You can create a smaller package off the original package, and include only those query subjects that you need in your dashboards and explorations. Enriching this smaller package requires less time and memory.
About this task
You can enrich a package metadata by using the automatic or manual process. The automatic process evaluates all query items of all selected query subjects in the package, and automatically applies the data characteristics to them. To minimize the impact on the system, you can deselect namespaces or individual query subjects to exclude them from the enrichment process.
In the manual process, you explicitly apply the data characteristics to individual query items. The manual process is not applicable for dimensional data.
When enriching a package, you typically start with the automatic process. Use the manual process to enrich only a small subset of query items or to override values that were set incorrectly by the automatic option.
The automatic enrichment includes the option to retrieve sample data. When this option is selected, the Cognos Analytics query engine connects to the data source and reads a sample of its data. The enrich dialog box allows the sample size to be changed. Setting the sample size to a low value, or not sampling at all, reduces the amount of information that the enrichment can gather. The amount of sampled data also depends on the signons that are used to access the package underlying data sources. An ideal signon can access the tables, views, and columns that the query subjects are based on, and a representative number of rows and values in the queried tables and views.
To access the Enrich package functionality, you need write permissions for the package.
- Locate the package or its shortcut in Team content or My content.
From the package or shortcut context menu , select Enrich package.
Tip: If a package was used as a data module source, you can enrich the package in the modeling user interface, from the Sources pane.
Select one of the following options.
- Enrich automatically
Most of the time, start with this option. The status information shows you the dates when the package was last published and enriched (if it was enriched before).
- In the Select tables panel, you can deselect the query subjects that you
don't want to be evaluated by the enrichment process. By default, all visible query subjects in the
package are evaluated.
This option gives you the opportunity to exclude the query subjects that aren’t used in your dashboards or explorations, and therefore reduce the time and memory usage by the system during the enrichment process.
- To enable data sampling, select the Retrieve sample data checkbox, and
specify the number of rows of data to be retrieved.
The data sample includes some deeper data characteristics that support the product functions that are behind the optimized user experience in dashboards, explorations, and other components. Extracting too many rows might impact the system performance. No data sampling, or too few rows might not provide enough information. Clearing this checkbox reduces the time and memory usage during the enrichment process, but the expected information might not be gathered.
For more information, see Data sampling.
- Click Run.
Depending on the number of query subjects involved, the enrichment process can take some time, potentially even hours. After the process is finished, an information message shows you the results of the process. Even if only a certain percentage of the query subjects were enriched, you might have enough data to support the AI-functions in your dashboards and explorations.
- Click Close.
- In the Select tables panel, you can deselect the query subjects that you don't want to be evaluated by the enrichment process. By default, all visible query subjects in the package are evaluated.
- Enrich manually
Use this option to enrich individual query items.
- Expand the package. Then, expand a query subject, and select one or more query items.
- From the Define data representation drop-down menu, select the option that you want the data in the query to represent, either Time or Geographic Location, and their specific values. The Default value allows to propagate settings from the source.
- Click OK.
- Enrich automatically