6、杜立-Flink SQL的功能扩展与深度优化-ZH.pdf

1、Flink SQL的功能扩展与深度优化 杜立 / 腾讯高级工程师 The Extensions And Optimizations For Flink SQL In Tencent Practice 窗口操作功能扩展 02 背景及现状 01 回撤流优化 03 未来规划 04 Background And Current SituationExtensions About Flink WindowOptimizations For Retraction MechanismFuture Plans 背景及现状01 Background And Current Situation 如何创建一个Fli

2、nk任务？ 基于底层API开发，面向开发人员 Based On Low-level API, For Developers 功能灵活多变 Most Flexible 易于性能调优 Easy To Optimize Performance 不便于扩展和升级 Expensive to expend or upgrade 学习门槛高 Highest learning threshold JARCanvasSQL How To Create A Flink Job? 基于页面拖拽开发，面向小白用户 Based On WEB UI, For Novice Users 功能比较灵活 More Flexib

3、le 较完整的功能覆盖 More Complete Functions Coverage 直观易懂 Intuitive and understandable 配置繁琐 Cumbersome Configurations 基于标准SQL语言，面向分析人员 Based Standard SQL, For Analysts 清晰简洁，易于阅读 Clear And Concise 易于升级 Easy To Upgrade Job 可重用性强 Easy To Reuse The Logic ？ Flink SQL当前的痛点 l 语法不够统一，有一定的学习成本 The SQL Syntax Is Not Uniform l 功能覆盖不够全 Insufficient Scenarios Coverage l 性能调优困难 Hard To Performance Tuning l 问题定位困难 Hard