Once files have been registered, you can create datasets and then set up relationships between datasets. However, if you use the Refine option when you register the file, you can avoid additional data processing overhead and may not need to create a dataset.
Kyvos supports building on cloud S3 and on-premise Hadoop.
People with administrator or data designer rights can create and use datasets.
Things to know before you start
Before you start creating datasets, you should already know some information about the raw data and about how it will be used. For example, you can use a dataset as source data for another dataset and apply transformations to the data when you set up the dataset.
You can specify information about the appearance of the data, such as the date format, currency format, and so on. Actions allow you to transform the data before you use it in analytics. For example, you can take different sources of data and specify that the fields be treated the same way so that you can use them together for analysis.
You can preview the data from any action or dataset by double-clicking the item and clicking Preview.
You can transform your data by choosing Utilities then Transform Data from the Toolbox. As you modify a dataset, some changes require a full build and some such as changing a filter only need an incremental build. See Impact of changes to datasets for details.
The major topics in this section are:
- Creating datasets
- Adding new fields
- Changing source data
- Transforming a dataset
- Transformation actions
- Filtering partitions
- Materializing a dataset
- Importing Java formulas
- Creating MDX queries
- Scheduling dataset builds
- Viewing dataset build summaries
- Impact of changes to datasets
- Using external tasks
- Using the Java Row Processor
- Using lookup (GIS)
- User-defined functions