Administration and Limitations

Auto Insights are on by default and run when you upload and save your dataset. You can refresh or disable them if required. There are some limitations on the datasets supported by Auto Insights in Oracle Analytics.

When Do Auto Insights Run and How Do You Refresh or Disable Them?

Auto Insights start running on datasets when the datasets are added to Oracle Analytics and saved. It typically takes a minute or two to complete before the Auto Insights icon turns yellow. The algorithms also recalculate every time the dataset structure is edited, for example, if you edit a column in the dataset (such as by changing the name, changing an aggregate rule, or hiding a column), or if you resave the dataset. You can manually initiate a refresh of Auto Insights in two ways:

Edit the settings in the Insights panel, and click Apply. This refreshes the most interesting insights with the new settings. See the previous paragraph.
Turn Auto Insights off and then back on in the dataset Inspect menu. This forces a complete refresh of the Auto Insights results and recalculates a new set of statistics on the dataset. This is a deeper refresh than the settings update suggested in the previous bullet and can take a few minutes to complete, depending on the size of the dataset.

For information about refreshing Auto Insights to include filter changes, see Refresh Auto Insights to Include Filters.

Auto Insights are turned on by default. The administrator controls the use of Auto Insights in your instance by setting an option in the Performance and Compatibility Options of the administration console. You can turn Auto Insights on and off for a dataset by displaying the Inspect menu and deselecting the Insights option on the first page of the dialog. Turning Auto Insights back on recalculates statistics on the dataset.

Supported Dataset Types and Known Limitations

Oracle Analytics generates insights for all datasets that contain less than 300 million cells (where the number of cells is calculated as the number of columns multiplied by the number of rows). Datasets larger than 300 million cells aren't currently supported. Additionally, the following types of datasets are unsupported:

Multidimensional sources such as Essbase, Oracle Planning and Budgeting Cloud Service, and Analytic Views
Subject Areas or Local Subject Areas

You can use multi-table datasets if you set the Preserve Grain option on the Fact Table in the model. Setting this option indicates which table acts as the fact table when multiple tables are joined in a model. Set the Preserve Grain option by right-clicking the table in the data model.

If you have a dataset of an unsupported type, you can generate an extract of these datasets and create a file-based or table-based dataset in Oracle Analytics, then run Auto Insights on them.

Auto Insights display text descriptions of insights and calculations in English.

After you refresh the Insights panel, if there are complex filters applied in the workbook, Auto Insights may not render certain types of visualizations.