Iterate over chunks of data with IJobChunk
To iterate through your data at the level of entire chunks of entities, implement IJobChunk inside a system
When you schedule an IJobChunk job in the OnUpdate method of a system, the system uses the entity query you pass to the schedule method to identify the chunks that it should pass to the job.
The job invokes your Execute method once for each matching chunk, and excludes those where no entities match the query because of enableable components. In your job’s Execute method, you can iterate over the data inside each chunk, entity by entity.
Note
Iterating with IJobChunk is more complicated and requires more code setup than using IJobEntity (which generates an IJobChunk under the hood). To automatically benefit from any future source-generation improvements, most jobs that perform a single iteration over a chunk's entities should prefer IJobEntity (optionally implementing the IJobEntityChunkBeginEnd interface as well, for any custom chunk-level operations before and after the core entity iteration loop).
Some examples of use cases which require IJobChunk:
- jobs which do not iterate over each chunk's entities at all (e.g. gathering per-chunk statistics)
- jobs which perform multiple iterations over a chunk's entities, or which iterate in a unusual order.
For more information, the ECS samples repository contains a simple HelloCube example that demonstrates how to use IJobChunk.