Traditional caching in Rock is limited to specific blocks, or to a particular format when using the Lava cache tag. Persisted Datasets are an always-ready cache that allow you to shape data for speed and use across many different blocks, and with different types of markup. Persisted Datasets are cached on the database or in memory using a job, so they’re quick every time. Persisted Datasets should be used when a large dataset is resource intensive to process, leaving people waiting seconds, or even minutes, for results. They can also be used when certain queries, like getting an attribute of another attribute, would cause issues at scale. Persisted Dataset Example {% assign data = 'mydataset' | PersistedDataset %} {%- for item in data -%} {{ item.Title }} {%- endfor -%} If you want more details, we have a whole chapter on Persisted Datasets.