Apache Tajo 0.2.0 发布,分布式数据仓库
- - 开源中国社区最新新闻Tajo 是一个分布式数据仓库系统,基于 Hadoop 实现,特点是低延迟、高可伸缩,提供专用查询和 ETL 工具. 完全分布式的 SQL 查询处理,基于存储雨 HDFS 的大数据集. 超低响应时间(约100毫秒),在合理数据范围内的简单查询. 容错支持,避免某些任务失败后的查询重启. 动态调度,处理和异构集群节点故障.
Tajo 是一个分布式数据仓库系统,基于 Hadoop 实现,特点是低延迟、高可伸缩,提供专用查询和 ETL 工具
特点:
Apache Tajo 0.2.0 包含如下改进:
* Add cost-based join optimization * Allow inline view use (i.e., table subquery) * Add various string functions, such as upper, lower, (L|R)TRIM, split_part, and regexp_replace. * Allow in predicate support * Improve significantly scan performance * Add INSERT OVERWRITE statement * Add CREATE TABLE statement * Add HiveQL mode * Allow configurable NULL character for CSVFile format * Allow compression/decompression of CSVFile (all codecs supported by Hadoop) * Add the extensible rewrite rule engine * Add tajo_dump, a backup and restore utility * Allow BETWEEN predicate * Add Tajo Resource Manager specialized for low-latency queries