Nesting a data package

Data packages on member nodes can exist as independent packages or in groups (nested data packages). Much like we can group multiple data files together with a common metadata file, we can group related data packages together with a common “parent” data package.

The structure of nested data packages resembles a pyramid. There is one top level, or “parent”, with one or more data packages, or “child packages”, nested beneath it. There is no limit to how many nested levels can be created, but packages do not generally exceed 3 levels. This “grandparent” has 5 child packages (nested datasets), all of which have child packages of their own.

Here are some common uses for nesting:

• collected data vary by year
• an NSF award funds several related projects
• data collection is still ongoing
• data files exceed the 1000 file limit per data package