OpenMLDB Weekly：重新设计和实现高性能单机版

内容简介：OpenMLDB Summary 本周合并 Pull requests 8个，新增Pull requests 5个，关闭 Issues 11个，新增 Issues 20个。总计84个文件修改，新增6677行代码，删除511行代码。 Merged Pull Requests feat: support spark.m...

OpenMLDB

Summary

本周合并 Pull requests 8个，新增Pull requests 5个，关闭 Issues 11个，新增 Issues 20个。总计84个文件修改，新增6677行代码，删除511行代码。 

Merged Pull Requests

• feat: support spark.master config to run job in yarn or local#466 
• feat: read openmldb git properties and set in return string#464
• feat: enable `TestWindowUnion` test#467
• feat: bump snakeyaml from 1.17 to 1.26 in /test/batch-test/openmldb-batch-test#416 
• feat: add openmldb jmh#367 
• test: modify cicd#465 
• feat: add src test cicd#452 
• fix: fix a bug when turn on window parallelization and skew optimization#438 

Open Pull Requests

• Fix import error of OpenMLDB Python lib when installed as zip format#439 
• fix: remove dup apply pass on the same physical op.#453 
• feat: ddlparser extracts indexes from sql#457 
• build: add deployment script#469 
• style: update hybridse header guard style#470 

Close Issues

• Support Spark local jobs in TaskManager for local OpenMLDB setup#456 
• Return the openmldb version string for Batch#463 
• Enable the unit test of TestWindowUnion#339 
• 创建索引，ttl_type=test，创建成功，期望返回错误信息#162 
• Deploy all-in-one java package#84 
• move demo to main repo#120 
• 创建absorlat类型的索引，desc展示不出来#163 
• scripts: package java sdk set cmake type to release#371 
• feat: support insert multiple rows into a table using a single SQL insert statement.#398
• Fix enable WindowSkewOpt and WindowComputeParallelOpt at the same time#444 
• add more status badges to README#313 

Open Issues

• Support passing multiple databases for parsing each SQL statement#468 
• Set index if there is no index info in create statement#462 
• Update catalog in stand alone mode#461 
• Add `SHOW DEPLOYMENT` and `DROP DEPLOYMENT` command#460 
• Recover metadata when nameserver restart in stand alone mode#459 
• Store metadata in system table#458 
• Support `SELECT INTO OUTFILE` syntax#455 
• ci: pre-build macos pkg needs to select sdk#454 
• Support querying data from Trino in oenmldb CLI#451 
• Add SHOW JOBS command for CLI and TaskManager#450 
• Support generating and storing job id for TaskManager#449 
• Integrate the ability of SQL analysis by trino in CLI#448 
• Add `DEPLOY` command#447 
• Write the result of select sql into file#446 
• Support `SET` syntax#445 
• Add `import` command#443 
• Support `engine` option in `create` statement#442 
• Use nameserver ip and port to connect openmldb#441 
• Support stand-alone mode#440 
• Fail query window+lastjoin when turn on `enable_batch_window_parallelization`#437 

Contributors

• aceforeverd (teapot@aceforeverd.com)
• Chen22 (jingchen2222@gmail.com)
• Kanekanekane (1290561498@qq.com)
• Stark (hujinlei1999@qq.com)
• tobe (tobeg3oogle@gmail.com)
• wuyou10206 (zw1020688@163.com) 

Highlights

本周在特性开发、缺陷修复、性能测试、CICD集成方案都有进一步完善。功能上在TaskManager上支持了单机批处理模式，Bug方案修复Git版本信息展示、批模式同时运行窗口倾斜优化以及窗口并行优化的问题，性能测试方面集成了openmldb-jmh模块加强对 Java 模块的性能回归测试，CICD集成方面加入了基于源码分支的SRC测试和基于特定版本的PKG测试，包含数千个case的大规模集成测试也可以在Github中手动出发运行。本周新增Issue较多，侧重于高性能单机版的重新设计和实现，对高性能数据库优化感兴趣的也可以关注一下。

欢迎更多开发者关注和参与OpenMLDB开源项目。

以上就是本文的全部内容，希望对大家的学习有所帮助，也希望大家多多支持 码农网

查看所有标签

本站部分资源来源于网络，本站转载出于传递更多信息之目的，版权归原作者或者来源机构所有，如转载稿涉及版权问题，请联系我们。