- All Implemented Interfaces:
- HudiSplitWeightProvider
public class SizeBasedSplitWeightProvider
extends Object
implements HudiSplitWeightProvider
Hudi split weight provider based on the split size.
The standard split size, `standardSplitSize` in bytes, corresponding split weight "1" and the minimum weight
`minimumWeight` are given during the initialization.
Given the split size in bytes as `splitSizeInBytes`, the raw weight is calculated as
`splitSizeInBytes / standardSplitSize`. If the weight is smaller than `minimumWeight`, `minimumWeight` is used.
If the weight is larger than 1, 1 is used as the weight. The split weight is always in the range of
[`minimumWeight`, 1].