Apache MRQL 0.9.6-incubating 发布了,该项目目前还处于 Apache 基金会的孵化状态。
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 Query Language) 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
新版本改进记录包括:
New Features
[MRQL-72] Add support for stream input on TCP sockets
[MRQL-79] Add support for incremental query processing
Bugs
[MRQL-67] Fix maven build errors for JDK 1.8
[MRQL-69] Allow escaped quotes in JSON strings
[MRQL-73] Set the max number of tasks in Spark mode
[MRQL-76] Change the mrql.flink shell script to work with Flink 0.9.0
[MRQL-78] Turn off Flink log INFO messages
[MRQL-81] Fix the coercion of persistent collections to in-memory bags
[MRQL-82] Fix the split size calculation in map-reduce mode
Improvements
[MRQL-70] Update Flink version to 0.8.1
[MRQL-74] Update Flink version to 0.9.0
[MRQL-75] Add support for Hama v0.7 on Yarn
[MRQL-77] Support single Flink jobs on YARN
[MRQL-80] Support Flink 0.10
[MRQL-84] Improve performance of Incremental MRQL
Tasks
[MRQL-68] GitHub integration for MRQL
[MRQL-71] Add google analytics tracking and bring the MRQL web site up-to-date
[MRQL-85] Prepare for the 0.9.6 release
下载地址:http://mrql.incubator.apache.org/downloads.html
来自:开源中国社区

