发布于 2015-11-24 23:31:26 | 166 次阅读 | 评论: 0 | 来源: 网友投递
Apache Lens 统一数据分析接口
Lens 提供了一个统一数据分析接口。通过提供一个跨多个数据存储的单一视图来实现数据分析任务切分,同时优化了执行的环境。无缝的集成 Hadoop 实现类似传统数据仓库的功能。
Apache Lens 2.4.0-beta 发布,此版本更新内容如下:
[LENS-113] - Document how to document a feature in contributor guide
[LENS-399] - TestRewriting test is failing with powermock incompatibility with jdk8
[LENS-607] - Test Failures on java8
[LENS-724] - Bump version in master to 2.4
[LENS-732] - exclude patchprocess directory from licence check.
[LENS-748] - Change version number to 2.4.0-beta-SNAPSHOT in master
[LENS-790] - UI Module failing if maven version is < 3.1
[LENS-802] - Merge master with current-release-line for 2.4
[LENS-803] - Bump version in master to 2.5
[LENS-804] - Verify Lens 2.4 issues
[LENS-805] - Update 2.4 release documentation
[LENS-806] - Update gpg keys in dist repo
[LENS-224] - FinishedLensQuery is not storing selectedDriver
[LENS-446] - server params should not be passed to session and drivers
[LENS-576] - Cannot retrieve query results on lens server restart
[LENS-734] - Query output files should have the query name
[LENS-745] - Drop partition is assigning latest=null if latest partition is being dropped
[LENS-756] - Example query on customer_city_name failing
[LENS-762] - Partition Created for querystats is not honouring "lens.server.statistics.warehouse.dir"
[LENS-765] - null Error Message on Lens Server logs
[LENS-774] - Unable to override "lens.server.session.expiry.service.interval.secs" because of spelling mistake
[LENS-775] - NPE on closing HadoopFileFormatter
[LENS-777] - Query remains in queue even when driver is free
[LENS-792] - StatusPoller logs are useless
[LENS-795] - Look ahead is failing with partition not found if any process time partitions are absent
[LENS-816] - ThreadSafeEstimatedQueryCollection is ignoring driver argument in getQueries and getQueryCount
[LENS-826] - lens.client.query.poll.interval needs doc update
[LENS-847] - Columnar JDBC Rewriter in incorrectly pushing filter against wrong alias when fact columns map to a single dimension
[LENS-852] - lens-ship-jars in current-release-line branch should have parent version 2.4.0-beta-SNAPSHOT instead of 2.5.0-beta-SNAPSHOT
[LENS-341] - Lens is not building against on JAVA-8
[LENS-392] - Look ahead timerange should not look for all finer granularity partitions to be present
[LENS-533] - Critical metrics for lens
[LENS-559] - Add inline values to be specified for XDimAttribute in xsd
[LENS-629] - A new, improved web client.
[LENS-693] - Queries get purged to DB as soon as they finish
[LENS-710] - Allow column name mapping for few/all columns in underlying storage tables
[LENS-716] - Support for !=, in and not in predicates in elastic search driver
[LENS-731] - Enhance cube.xsd to accept Hierarchical dimattribute and inline dimattribute
[LENS-737] - Throw single error out with LensMultiException
[LENS-739] - Generate udf jar for lens-regression
[LENS-753] - Queue number for queries
[LENS-755] - rename LogicalTableCrudCommand to ConceptualTableCrudCommand and PhysicalTableCrudCommand to LogicalTableCrudCommand
[LENS-757] - Examples should continue to run in case submission failures
[LENS-772] - Both time and size based rolling for logs
[LENS-782] - UI support for the saved, parametrized query feature.
[LENS-796] - Slow response times for /metastore/nativetables API
[LENS-797] - Improve CLI documentation to mention the helpful role of pressing tab while writing commands
[LENS-831] - Create a single jar for shipping with hive queries
[LENS-742] - Saved query and parameterization
[LENS-752] - Support flattening of columns selected through bridge-tables(many-to-many relationships)
[LENS-108] - Improve developer documentation
[LENS-187] - Move cube specific error message codes from InMobi's Hive code to Lens
[LENS-758] - Remove feather logo from the repository, use already existing image to display on website
[LENS-784] - setup build job with java8 on jenkins
[LENS-801] - Release Lens 2.4.0-beta
更多内容请看发行说明。
Lens 提供了一个统一数据分析接口。通过提供一个跨多个数据存储的单一视图来实现数据分析任务切分,同时优化了执行的环境。无缝的集成 Hadoop 实现类似传统数据仓库的功能。
该项目主要特性:
简单元数据层为数据存储提供抽象视图层
单一的共享模式服务器,基于 Hive 元存储。模式通过数据管道 HCatalog 和分析应用进行共享:
OLAP Cube QL 类似 SQL 的高级语言用来查询和描述存放在不同数据立方体 (Cubes) 中的数据集
JDBC 驱动和 Java 客户端库来处理查询
Lens 应用服务器 - 这是一个 REST 服务器允许用户查询数据,更改数据模型,调度查询和查询的配额限制
基于驱动的架构 允许在报表系统中进行嵌入,例如 Hive、列数据存储、Redshift 等
基于成本算法的引擎选择 - 该算法可优化资源的使用,通过对查询的复杂度自动选择最佳执行引擎
Apache Lens 的架构如下: