TiSpark 1.0.1

• Fix unsigned index
• Compatible with TiDB before and since 48a42f

TiSpark 1.0 GA

TiSpark provides distributed computing of TiDB data using Apache Spark.

• Provide a gRPC communication framework to read data from TiKV
• Provide encoding and decoding of TiKV component data and communication protocol
• Provide calculation pushdown, which includes:
  • Aggregate pushdown
  • Predicate pushdown
  • TopN pushdown
  • Limit pushdown
• Provide index related support
  • Transform predicate into Region key range or secondary index
  • Optimize Index Only queries
  • Adaptive downgrade index scan to table scan per region
• Provide cost-based optimization
  • Support statistics
  • Select index
  • Estimate broadcast table cost
• Provide support for multiple Spark interfaces
  • Support Spark Shell
  • Support ThriftServer/JDBC
  • Support Spark-SQL interaction
  • Support PySpark Shell
  • Support SparkR

1.0-RC2

1. Improve statistics info management and plan cost evaluation.
2. Index read will be downgraded to table scan if index read causes too much pressure(estimated) on TiKV.
3. Improve TiSpark error handling logic.
4. Force refresh metadata support.
5. Improve expression system.
6. Fix unsigned datatype used in index.
7. Minor bug fixes.

1.0-RC1

TiDBMapDatabase fails when Decimal length exceeding 38 (#285)