95. A quick note about the performance of Java Mapping Framework (Baeldung)
According to [1], Creating large Java applications composed of multiple layers require using multiple models such as persistence model, domain model or so-called DTOs. Using multiple models for different application layers will require us to provide a way of mapping between beans.
Doing this manually can quickly create much boilerplate code and consume a lot of time. Luckily for us, there are multiple object mapping frameworks for Java.[1]
This article just take a quick note on the performance of Java Mapping frameworks, include MapStruct, ModelMapper, JMapper.
Benmarks
The average running time per operation (in ms), the less is better:
| Framework Name | Average Running Time (ms per operation) |
|---|---|
| MapStruct | 10⁻⁴ |
| JMapper | 10⁻⁴ |
| Orika | 0.007 |
| ModelMapper | 0.137 |
| Dozer | 0.145 |
Throughput benmarks: the benchmark returns the number of operations per second (the more is better)
| Framework Name | Throughput (operations per ms) |
|---|---|
| JMapper | 3205 |
| MapStruct | 3467 |
| Orika | 121 |
| ModelMapper | 7 |
| Dozer | 6.342 |
Personally, I only had a chance to work with MapStruct and ModelMapper. Not working with other mapping framework.
Conclusion:
- MapStruct is far faster than ModelMapper
- MapStruct use compile-time code generation. The generation code is nearly the same way we write the convertion code manually.
- ModelMapper use reflection, less performance.
- MapStruct is more explicit when define mapping. ModelMapping provide more “intelligent” mapping option, but it can be hard to debug and understand. I’m not prefer it
- Both support nested object mapping