The current dictionary encoder implementation (org.apache.arrow.vector.dictionary.DictionaryEncoder) has heavy performance overhead, which prevents it from being useful in practice:
- There are repeated conversions between Java objects and bytes (e.g. vector.getObject).
- Unnecessary memory copy (the vector data must be copied to the hash table).
- The hash table cannot be reused for encoding multiple vectors (other data structure & results cannot be reused either).
- The output vector should not be created/managed by the encoder (just like in the out-of-place sorter)
- The hash table requires that the hashCode & equals methods be implemented appropriately, but this is not guaranteed.
We plan to implement a new one in the algorithm module, and gradually deprecate the current one.