Apache MRQL 0.9.4-incubating 发布
3月21日 深圳 OSC 源创会开始报名罗，送华为海思开发板Apache MRQL 0.9.4-incubating 发布，此版本是第二个发布的版本，现已提供下载：
http://www.apache.org/dyn/closer.cgi/incubator/mrql
更新内容：
新特性
[*] Add support for Apache Flink
[*] Add Flink support info on the project web site
[*] Add support for MRQL streaming in spark streaming mode
Bugs 修复
[*]] Post-release improvements
[*] Fix bug in XML projections
[*] Fix the map-reduce join for a complex reducer on map-reduce mode
[*] MRQL does not build with Java 8
[*] Fix various bugs for the Flink evaluation mode
[*] Fix global variable bindings in Spark and Flink modes
[*] Fix result dumping to a text file in Flink mode
[*] Support for Spark 1.1.0
[*] Fix Flink evaluation mode to pass all Junit tests
[*] Fix processing of directories of text files in MapReduce mode
[*] Fix config file and shell run scripts to support multiple HDFS users
[*] Fix building dependencies for hadoop 1.x
[*] Fix bug on select over union queries
[*] Fix Map-Reduce mode on MultipleInput sources
[*] Fix maven build errors for JDK 1.8
改进
[*] Add support for tracing
[*] Adjust the split size of a map-reduce input file based on the number of requested nodes
[*] Add support for Hadoop Sequence input format in flink mode
[*] Improve total aggregations and repetitions with shared results in Flink mode
[*] Make floating-point literals to be double instead of float
任务
[*] Prepare for 0.9.4 release
发行说明：
http://mrql.incubator.apache.org/ReleaseNotes-0.9.4.html
MRQL (发音 miracle) 是一个查询处理和优化系统，适用于大规模分布式的数据分析。MRQL (MapReduce QueryLanguage) 是一个在计算机集群中对大规模数据的类 SQL 查询语言。MRQL 查询处理系统可使用如下三种模式评估 MRQL 查询：
[*]使用 Hadoop 的 Map-Reduce 模式
[*]使用 Apache Hama 的 BSP 模式 (Bulk Synchronous Parallel mode)
[*]基于 Apache Spark 的 Spark 模式
页:
[1]