Some key features and improvements in this release:
1. Supported carbon as Spark datasource using Spark's
2. Supported Spark 2.3.2 version
3. Supported Hadoop 3.1.1 version
4. Improved compression and performance of non dictionary columns by
applying an adaptive encoding to them,
5. Supported MAP datatype in carbon.
6. Supported ZSTD compression for carbondata files.
7. Supported C++ interfaces to read carbon through SDK API.
8. Supported CLI tool for data summary and debug purpose.
9. Supported BYTE and FLOAT datatypes in carbon.
10. Limited min/max for large text by introducing a configurable limit
to avoid file size bloat up.
11. Introduced multithread write API in SDK to speed up loading and
12. Supported min/max stats for stream row format to improve query
13. Many Bug fixes and stabilized carbondata.