2.3 Open and Closed Sources

When a source is created, BigML tries automatically to figure out how to extract rows from the raw data. For instance, if the file is a CSV file, most of time BigML will discover what columns and column types it has, and they will become the source’s fields.

Occasionally however, if BigML doesn’t get everything right, users can edit that metadata and adjust the definitions of the source’s fields until they are correct. Source is the only BigML resource that is initially mutable.

A source is open when it can be modified. All sources are initially created open.

A source can be changed from open to closed. For composite sources, this is done by performing the Close This Composite Source operation.

For a non-composite source (either single or inline source), it will be automatically closed when its dataset is created. When creating a dataset from a composite source, the composite source is automatically closed too.

For a composite source, the lock in its icon under “Type” column represents whether it is open or closed, as seen in the figure above. The open or closed status is also in the source’s description string.

Once a source has been closed, it cannot be modified anymore.

What if you want to change something in a closed source? You can clone it. Cloning a source creates a new and open source that shares the original underlying data. Cloning is quick and efficient. Because a cloned source is created as an open source, it can be modified again.

An open source can be cloned too.