The Data API

The unified Data API aims to provide a consistent way of accessing and modifying data. ‘Data’, in this context means any data that is consistently synchronised between client and server. It can be changed server-side and then those changes will be synchronised to the connected clients. This includes, among many others, the text on a sign, the looks of a horse or the health of any living entity.

Where other approaches define the available data using interfaces and inheritance (like a LivingEntity interface providing getter and setter functions for current and maximum health), in Sponge every entity, block etc. is completely oblivious to what data it holds. While this may appear less straightforward than direct accessor methods, it is foremost far more extensible. And thanks to the addition of Keys, simply accessing specific values is no less straightforward.


As of writing, a few parts of the Data API are not implemented. If you are trying to access an API and are receiving an empty Optional when one is not expected, refer to the Implementation Tracker, ask in the #spongedev IRC channel or on the Forums to find out if the data you need to work with is available yet.


On first glance at the API docs, the data API threatens to overwhelm you with lots of interfaces and packages. But to simply use the data API, you will not have to deal with many of them, as most interfaces found there are just specific data manipulators.


A data holder is just that - something that holds data. It provides methods to retrieve and offer back data. The interface itself is completely oblivious to the type of data held. Since only the implementations will possess this knowledge, it is possible to ask a DataHolder to provide data it does not have or to accept data it cannot use. In those cases, the return values of the methods will provide the information that data is not available (via Optional.empty()) or not accepted (via the DataTransactionResult).


A property too is data, but not synchronised between server and clients. Therefore, it can only be changed by modifications present on both client and server. Since Sponge is not intended to require a client-side counterpart, properties are not modifiable. Examples of properties are the harvesting abilities on tools (represented as HarvestingProperty or the damage absorption of an equippable armour item (represented as DamageAbsorptionProperty).


A data manipulator represents points of cohesive data that describes a certain component of its holder. For example HealthData, which contains both current and maximum health. If a data holder has HealthData, it has health that can somehow be depleted and replenished and can die if that health is depleted. This allows for the re-use of such components over the API and prevents duplication of accessor methods. For example, sheep, stained glass blocks and leather armour all can share the DyeableData holding the colour they are dyed in.


A Key is a unique identifier for a single point of data and can be used to directly read or set that point of data without worrying about data manipulators. It was designed to provide a convenient way of accessing data similar to direct getter/setter methods. All keys used within Sponge are listed as constants in the Keys utility class.


Within the Data API, a value referred to by a Key is encoded in a container object. For this documentation, it is referred to as ‘keyed value’ to avoid confusion with the actual value. A keyed value encapsulates the actual data value (if it is present), a default value (to be used if no direct value is present) and the Key by which the value is identified.