Yuming Wang

```shell bin/spark-shell --conf spark.jars.repositories=https://mmlspark.azureedge.net/maven --packages com.microsoft.ml.spark:mmlspark_2.11:1.0.0-rc3 ```

@Yikf Is there any mainstream database that supports this optimization?

I think it's by design. So enabling `spark.sql.cbo.enabled` is what you want?

cc @cloud-fan

Part of the TPC-DS q24a query plan. Before this PR | After this PR -- | --  | 

@cwang9208 Could you add my WeChat: yumwang666?

@sigmod

@cloud-fan

cc @cloud-fan @sunchao

cc @HyukjinKwon @cloud-fan