问题
It's quite simple to use row denormaliser to achieve pivots when we have few records which can be written manually in denorm step,but what when there's hundreds of thousands of records? I tried using etl metadata injection step, but I was unable to achieve my desired output.
Here is the link to my previous post where my source data has been defined.
and here is what I have tried
row denormaliser options
pivot_inject_etl_metadata.ktr
回答1:
Try something like this:
pivot_inject_etl_metadata.ktr:
Group by: sku_id
Add constants:
ETL Metadata Injection:
来源:https://stackoverflow.com/questions/35715048/pivots-using-etl-metadata-injection