Samples

Some datasets include a 1% data sample, which supports faster exploratory queries for very large tables. Datasets can also have fewer access restrictions to the sample than the full dataset.

Methodology

Sampling is applied independently to each version of a dataset. Samples can either be a random sample applied to all tables in the dataset, or a sample created on a variable which will be applied to all tables in the dataset that contain that variable.

Usage

You can see if a sample exists on a dataset and how the sample was created in the right bar next to the list of tables. The tables list can also be filtered to see which tables have a sample. If you only have sample access to this dataset you might see a Sample tag next to table names, which means you are looking at the sampled table of data.

When adding a dataset to a workflow, the dataset will always default to the sample if one exists. A dataset can be switched between the sample and the full data at any time.

For editors: How to create a sample