IncrementalDataset 实现了一种增量数据处理的能力,基于了PartitionedDataset 同时包含了checkpoint 确保数据处理的准确性,对于
checkpoint 可以配置自己的函数
参考定义
- 参考catalog 定义
my_partitioned_dataset:
type: partitions.IncrementalDataset
path: s3://my-bucket-name/path/to/folder
dataset: pandas.CSVDataset
checkpoint:
# update the filepath and load_args, but keep the dataset type unchanged
filepath: gcs://other-bucket/CHECKPOINT
load_args:
k1: v1
- 自定义函数的catalog 配置
my_partitioned_dataset:
type: partitions.IncrementalDataset
path: s3://my-bucket-name/path/to/folder
dataset: pandas.CSVDataset
checkpoint:
comparison_func: my_module.path.to.custom_comparison_function # the path must be importable
- 指定checkpoint 具体文件
my_partitioned_dataset:
type: partitions.IncrementalDataset
path: s3://my-bucket-name/path/to/folder
dataset: pandas.CSVDataset
checkpoint:
force_checkpoint: 2020-01-01/data.csv
说明
增量能力还是很重要的,尤其是在数据量比较大的场景,使用好IncrementalDataset 可以提升数据处理速度
参考资料
https://docs.kedro.org/en/stable/data/partitioned_and_incremental_datasets.html
标签:IncrementalDataset,type,dataset,checkpoint,kedro,简单,path,my From: https://www.cnblogs.com/rongfengliang/p/18353699