DataSourceV2 should expose row count and attribute statistics

During investigation of CBO and DataSourceV2 we found, that

org.apache.spark.sql.sources.v2.reader.Statistics

misses attribute/column statistics and that

DataSourceV2Relation#computeStats

wraps

org.apache.spark.sql.sources.v2.reader.Statistics

into

org.apache.spark.sql.catalyst.plans.logical.Statistics

without forwarding the optional rowCount if present.

However rowCount and attributeStats are used during CBO e.g. in JoinEstimation and AggregateEstimation.

We propose that:

• org.apache.spark.sql.sources.v2.reader.Statistics mirrors org.apache.spark.sql.catalyst.plans.logical.Statistics
• DataSourceV2Relation forwards all the information to be available during CBO