Add renaming transformer

[yruslan]

Background

The idea of the transformer was suggested by @kevinwallimann after a discussion on how Hive handles parquet partition directories. Since Hive (and most SQL in general) is case insensitive, and HDFS is not, Hive expects a convention that all partitioning columns will be lowercase. This is configurable on the global Hive level, but end-users of Hive don't have always permissions to change these configs. So making partitioning columns lowercase seems like the most workable solution for this particular use case.

But, let's say, have a data source in Kafka that has a column name 'infoDate'. If we want to partition by this coolumn, we can specify this in Hyperdrive:

writer.parquet.partition.columns=infoDate

But this will create a folder structure like this:

/bigdata/somedata/infoDate=2020-10-01/
/bigdata/somedata/infoDate=2020-10-02/
...

even if we specify writer.parquet.partition.columns=infodate, the partitioning column will be as defined in the data source since the default transformer doesn't change column names in any way.

Feature

Implement a renaming transformer in which a user can specify one or more columns to rename.