The dataset Go package is still rapidly evolving though it is now commonly used in Caltech Library. As a result of this evolution we are experimenting with two different file layouts currentely. The older layout is called “buckets”, the newer layout is a “pairtree”. Currently “buckets” is the default but you can migrate your collection form one to the other. Below is an example of migrating to the “pairtree” file layout.
# Migrating to a pairtree layout
dataset migrate mycollection.ds pairtree
To migrate from a “pairtree” to a bucket follows the same process.
# Migrating to a bucket layout
dataset migrate mycollection.ds buckets