Version 1.3.0
Spark-csv 1.3.0 adds following
Features:
- Printing content of records when failing or dropping
- Support for Timestamps in type inference
- Support for roundtrip null values of any type (by @andy327)
- Parsing double and float according to Locale if default cast fails (by @gasparms)
- Support for nullable quote character (by @jamesblau)
- Using system line separator for parsing (by Glenn Murray)
- Support for pruned scan for required fields (by @HyukjinKwon)
- Support for compressionCodec in all languages (by @msperlich)
- Support for data source short name with Spark 1.5+
Improvements
- Several documentation updates and fixes (by multiple community contributors)
- Several code style and compatibility automated tests (by @JoshRosen)
- Unit test consolidation (by @brkyvz)