Concept: Datasets
A dataset is a Prism Analytics object that controls some underlying data and describes some
processing logic to manipulate the data. A dataset is a description of the data,
otherwise known as metadata. It contains all of the data about the data, plus a subset
of example rows to help you understand the data values. You create datasets to prepare
data for analysis.
A dataset can describe either Workday or non-Workday (external) data. You might want to create
a dataset using external data to blend, transform, and enrich it with Workday data. This
enables you to analyze your Workday and non-Workday data together without having to
export it into a separate electronic data warehouse and business intelligence (BI)
application.
A dataset can also describe data that's output from another table or dataset by deriving the
dataset from an existing table or dataset.
You can create a:
- Base dataset. A base dataset is a dataset that is based on 1 or more source files. The source files can come from external sources (non-Workday data) or from the output of a Workday custom report. You create a base dataset when you create a new dataset using these options:
- from File. This option creates a dataset with external data that you upload in the browser.
- from SFTP. This option creates a dataset with external data that Workday retrieves from an SFTP server using an integration.
- from Custom Report. This option creates a dataset with Workday data.
- Derived dataset. A derived dataset is a dataset that is based on 1 or more existing tables or datasets. The source data of a derived dataset comes from the output of existing tables and datasets. You use derived datasets to blend data together from different sources, such as Workday data and non-Workday data. Some stage types, such as Join and Union, are only available to derived datasets. You create a derived dataset when you create a new dataset using theDerived Datasetoption.