原书出处
Data Management at Scale
11 Turning Data into Value
I recommend assembling your data pipelines as a series of isolated, immutable transformations that can be reused and combined easily. Input, transformation logic, and output in this case must be clearly separated from each other. For reproducible output data, I recommend versioning all pipelines. For better performance that takes advantage of the elasticity of modern infrastructure, you might want to engineer your pipelines to run
原书定位详情
来源:a19c7933-4eda-4176-aebd-30fe7e1ad292
正文位置:extracted/Data Management at Scale/_epub-src/OEBPS/ch11.html
EPUB SHA-256:48aa9e8e81be8a24243cd3a468bf103704d3a46e5c99edaa90b7433f7c941043