当前位置:网站首页>Flink learning 6: programming model
Flink learning 6: programming model
2022-07-04 04:18:00 【hzp666】

1. Interface layering
Flink Different levels of interfaces are provided , At the bottom is state data flow processing , The highest level is SQL

1.1 State data flow processing
Through the process function , Inherited to DataStream API in

Stateful data stream processing interface ,
1.2 DataStream and DataSet API
The core API
DataStream Used to deal with bounded and Unbounded data flow .
DataSet API Used to deal with bounded Data flow .
This interface provides a large number of common components

in addition ,
DataStream Integrate lower level functions
DataSet API Provides additional support for bounded data

1.3 Table API
1. Take the watch as the center , A two-dimensional table similar to a relational database

2. Provide comparable operations ,select, group by etc.

3.Table API You can customize functions to extend , But in terms of expression ability , Not as good as the core API.( however Table API The code is simpler )
4.Table API Is to specify what the program should do , Instead of specifying the exact execution steps of the program
5.Table API It will be optimized before implementation , in addition Table API and DataStream , DataSet Seamless switching between . So in the program, you can mix only .
1.4 SQL
Flink The most advanced encapsulation interface is flink SQL
Grammar and expressive ability are similar to Table API similar , The only difference is through the use of SQL Query language implementation program .
SQL Query and Table API Close interaction ,SQL Can be directly in Table API Query on the table created in

边栏推荐
- Evolution of MySQL database architecture
- Brief explanation of depth first search (with basic questions)
- Pytest multi process / multi thread execution test case
- Mitsubishi M70 macro variable reading Mitsubishi M80 public variable acquisition Mitsubishi CNC variable reading acquisition Mitsubishi CNC remote tool compensation Mitsubishi machine tool online tool
- 新型数据中心,助力加快构建以数据为关键要素的数字经济
- Getting started with the go language is simple: go implements the Caesar password
- Flink学习6:编程模型
- Global exposure and roller shutter exposure of industrial cameras
- 【罗技】m720
- The difference between bagging and boosting in machine learning
猜你喜欢

MySQL maxscale realizes read-write separation

mysql数据库的存储

leetcode刷题:二叉树09(二叉树的最小深度)

leetcode刷题:二叉树06(对称二叉树)

02 ls 命令的具体实现

Sales management system of lightweight enterprises based on PHP

10 reasons for not choosing to use free virtual hosts

The difference between bagging and boosting in machine learning

EV6 helps the product matrix, and Kia is making efforts in the high-end market. The global sales target in 2022 is 3.15 million?

【罗技】m720
随机推荐
SQL statement strengthening exercise (MySQL 8.0 as an example)
Why is the probability of pod increasing after IPtable
Flink学习7:应用程序结构
Illustrated network: what is the hot backup router protocol HSRP?
laravel admin里百度编辑器自定义路径和文件名
拼夕夕二面:说说布隆过滤器与布谷鸟过滤器?应用场景?我懵了。。
Unity移动端游戏性能优化简谱之 画面表现与GPU压力的权衡
【CSRF-01】跨站请求伪造漏洞基础原理及攻防
leetcode刷题:二叉树06(对称二叉树)
深度优先搜索简要讲解(附带基础题)
ctf-pikachu-XSS
华为云鲲鹏工程师培训(广西大学)
The difference between bagging and boosting in machine learning
Katalon uses script to query list size
毕业总结
Epidemic strikes -- Thinking about telecommuting | community essay solicitation
How was my life in 2021
Wechat official account web page authorization
ctf-pikachu-CSRF
10 reasons for not choosing to use free virtual hosts