Re: Issue of Hive parquet partitioned table schema mismatch

[Michael Armbrust]

>
> We have tried schema merging feature, but it's too slow, there're hundreds
> of partitions.
>
Which version of Spark?