当前位置:网站首页>2022-08-04 Select clause for clickhouse
2022-08-04 Select clause for clickhouse
2022-08-05 13:33:00 【Emperor Zun Wu Shi】
select clause:
SELECTThe clause determines which column fields or expressions will be returned by a query statementstyle.Unlike intuitive feelings, althoughSELECTLocatedSQLStart of statement, butIt is executed after the above clauses.After the other clauses are executed, SELECTApplies the selected field or expression to each row of data.
Join performance:
In order to optimizeJOINQuery performance, firstShould follow the principle of big on the left and small on the right , which isPlace tables with small amounts of data on the right.This is because executingJOINWhen querying, regardless of theWhich connection method is used, the right table will all be loaded into memory for comparison with the left table.Secondly,JOINQuery currently has no cache support , which means every timeJOINCheckQuery, even if the same is executed continuouslySQL, will also generate a new execution plan.If the application will be heavily usedJOINInquiry, you need to further consider using the upper applicationside caching service or useJOINTable engine to improve performance.Finally,If it is a query scenario where a large number of dimension attributes are completed, it is recommended to use the wordDictionary insteadJOINQuery .Because when a multi-table join query is performed, the query will be converted into two pairsconnected form, this“Snowball”-style queries are likely to cause performance problems.
Memory Limits
By default, ClickHouse uses the Hash Join algorithm.ClickHouse takes and creates a hash table for it in RAM.After a certain memory consumption threshold, ClickHouse falls back to the merge join algorithm.
If you need to limit the memory consumption of join operations, use the following settings:
- max_rows_in_join — Limits number of rows in thehash table.
- max_bytes_in_join — Limits size of the hash table.
When any of these limits are reached, ClickHouse operates as join_overflow_mode Setup instructions.
边栏推荐
- 115. In-depth explanation of the technical implementation of configuring the local SAP UI5 application to the local Fiori Launchpad
- 第一章 调度系统架构设计之线程池创建
- 小米Cyberdog源码开源啦!
- 机器学习:贝叶斯网络
- 记一次 .NET 某工控自动化控制系统 卡死分析
- pandas连接oracle数据库并拉取表中数据到dataframe中、筛选当前时间(sysdate)到一分钟之前的所有数据(筛选一分钟之内的数据)
- JS—动画
- 《MySQL核心知识》第5章:字段长度与类型宽度
- 适配器模式
- Horizon First Experience. Part 1
猜你喜欢
随机推荐
并发刺客(False Sharing)——并发程序的隐藏杀手
EMQ & IoTDB 联合 Meetup 杭州站 | 8月13日线下场报名开启
选择商城小程序源码的三个技巧!
怎么在微信上做小程序卖货?怎样做小程序卖货
在线问题反馈模块实战(十九):实现数据批量导出到excel文件中功能
VINS-Mono结果展示
2022-08-03 顾宇佳 学习笔记
内存管理架构及虚拟地址空间布局
《MySQL核心知识》第2章:MySQL常用的命令
fmincon函数 源代码
伙伴分配器的内核实现
配置网络源仓库
链表面试题-刷题
【搜索框】通用测试用例
SharePreference与MMKV对比
地平线初体验.上
行云管家荣获第十一届中国财经峰会“2022杰出品牌形象奖”
PySpark的存储不同格式文件
和为S的两个数字
Redis5.0的安装和配置









