当前位置:   article > 正文

Clickhouse 亿级数据性能测试_clickhouse单表最大行数

clickhouse单表最大行数

最近由于项目需求使用到了 clickhouse 做分析数据库,于是用测试环境做了一个单表 6 亿数据量的性能测试,记录一下测试结果,有做超大数据量分析技术选型需求的朋友可以参考下。

服务器信息

    • CPU:Intel Xeon Gold 6240 @ 8x 2.594GHz

    • 内存:32G

    • 系统:CentOS 7.6

    • Linux内核版本:3.10.0

    • 磁盘类型:机械硬盘

    • 文件系统:ext4

Clickhouse信息

    • 部署方式:单机部署

    • 版本:20.8.11.17

测试情况

测试数据和测试方法来自 clickshouse 官方的 Star Schema Benchmark,URL:https://clickhouse.com/docs/en/getting-started/example-datasets/star-schema/

按照官方指导造出了测试数据之后,先看一下数据量和空间占用情况。

数据量和空间占用

可以看到 clickhouse 的压缩率很高,压缩率都在 50 以上,基本可以达到 70 左右。数据体积的减小可以非常有效的减少磁盘空间占用、提高 I/O 性能,这对整体查询性能的提升非常有效。

supplier、customer、part、lineorder 为一个简单的「供应商-客户-订单-地区」的星型模型,lineorder_flat 为根据这个星型模型数据关系合并的大宽表,所有分析都直接在这张大宽表中执行,减少不必要的表关联,符合我们实际工作中的分析建表逻辑。

以下性能测试的所有分析 SQL 都在这张大宽表中运行,未进行表关联查询。

查询性能测试详情

Query 1.1

SELECT sum(LO_EXTENDEDPRICE * LO_DISCOUNT) AS revenueFROM lineorder_flatWHERE (toYear(LO_ORDERDATE) = 1993) AND ((LO_DISCOUNT >= 1) AND (LO_DISCOUNT <= 3)) AND (LO_QUANTITY < 25)
┌────────revenue─┐44652567249651└────────────────┘
1 rows in set. Elapsed: 0.242 sec. Processed 91.01 million rows, 728.06 MB (375.91 million rows/s., 3.01 GB/s.)

描行数:91,010,000 大约9100万

耗时(秒):0.242

查询列数:2

结果行数:1

Query 1.2

SELECT sum(LO_EXTENDEDPRICE * LO_DISCOUNT) AS revenueFROM lineorder_flatWHERE (toYYYYMM(LO_ORDERDATE) = 199401) AND ((LO_DISCOUNT >= 4) AND (LO_DISCOUNT <= 6)) AND ((LO_QUANTITY >= 26) AND (LO_QUANTITY <= 35))
┌───────revenue─┐9624332170119└───────────────┘
1 rows in set. Elapsed: 0.040 sec. Processed 7.75 million rows, 61.96 MB (191.44 million rows/s., 1.53 GB/s.)

描行数:7,750,000 775万

耗时(秒):0.040

查询列数:2

返回行数:1

Query 2.1

SELECT     sum(LO_REVENUE),    toYear(LO_ORDERDATE) AS year,    P_BRANDFROM lineorder_flatWHERE (P_CATEGORY = 'MFGR#12') AND (S_REGION = 'AMERICA')GROUP BY     year,    P_BRANDORDER BY     year ASC,    P_BRAND ASC
┌─sum(LO_REVENUE)─┬─year─┬─P_BRAND───┐644200056181992 │ MFGR#121633893460961992 │ MFGR#1210│     ........... │ .... │ ..........│396798929151998 │ MFGR#128353005130831998 │ MFGR#129└─────────────────┴──────┴───────────┘
280 rows in set. Elapsed: 8.558 sec. Processed 600.04 million rows, 6.20 GB (70.11 million rows/s., 725.04 MB/s.)

扫描行数:600,040,000 大约6亿

耗时(秒):8.558

查询列数:3

结果行数:280

Query 2.2

​​​​​​​

SELECT     sum(LO_REVENUE),    toYear(LO_ORDERDATE) AS year,    P_BRANDFROM lineorder_flatWHERE ((P_BRAND >= 'MFGR#2221') AND (P_BRAND <= 'MFGR#2228')) AND (S_REGION = 'ASIA')GROUP BY     year,    P_BRANDORDER BY     year ASC,    P_BRAND ASC
┌─sum(LO_REVENUE)─┬─year─┬─P_BRAND───┐664503494381992 │ MFGR#2221654232643121992 │ MFGR#2222│     ........... │ .... │ ......... │399075452391998 │ MFGR#2227406542018401998 │ MFGR#2228└─────────────────┴──────┴───────────┘
56 rows in set. Elapsed: 1.242 sec. Processed 600.04 million rows, 5.60 GB (482.97 million rows/s., 4.51 GB/s.)

扫描行数:600,040,000 大约6亿

耗时(秒):1.242

查询列数:3

结果行数:56

Query 3.1

​​​​​​​

SELECT     C_NATION,    S_NATION,    toYear(LO_ORDERDATE) AS year,    sum(LO_REVENUE) AS revenueFROM lineorder_flatWHERE (C_REGION = 'ASIA') AND (S_REGION = 'ASIA') AND (year >= 1992) AND (year <= 1997)GROUP BY     C_NATION,    S_NATION,    yearORDER BY     year ASC,    revenue DESC
┌─C_NATION──┬─S_NATION──┬─year─┬──────revenue─┐│ INDIA     │ INDIA     │ 1992537778456208│ INDONESIA │ INDIA     │ 1992536684093041│ .....     │ .......   │ .... │ ............ ││ CHINA     │ CHINA     │ 1997525562838002│ JAPAN     │ VIETNAM   │ 1997525495763677└───────────┴───────────┴──────┴──────────────┘
150 rows in set. Elapsed: 3.533 sec. Processed 546.67 million rows, 5.48 GB (154.72 million rows/s., 1.55 GB/s.)

 

扫描行数:546,670,000 大约5亿4千多万

耗时(秒):3.533

查询列数:4

结果行数:150

Query 3.2

​​​​​​​

SELECT     C_CITY,    S_CITY,    toYear(LO_ORDERDATE) AS year,    sum(LO_REVENUE) AS revenueFROM lineorder_flatWHERE (C_NATION = 'UNITED STATES') AND (S_NATION = 'UNITED STATES') AND (year >= 1992) AND (year <= 1997)GROUP BY     C_CITY,    S_CITY,    yearORDER BY     year ASC,    revenue DESC
┌─C_CITY─────┬─S_CITY─────┬─year─┬────revenue─┐│ UNITED ST6 │ UNITED ST619925694246807│ UNITED ST0 │ UNITED ST019925676049026│ .......... │ .......... │ .... │ .......... ││ UNITED ST9 │ UNITED ST919974836163349│ UNITED ST9 │ UNITED ST519974769919410└────────────┴────────────┴──────┴────────────┘
600 rows in set. Elapsed: 1.000 sec. Processed 546.67 million rows, 5.56 GB (546.59 million rows/s., 5.56 GB/s.)

查询列数:4

结果行数:600

Query 4.1

​​​​​​​

SELECT     toYear(LO_ORDERDATE) AS year,    C_NATION,    sum(LO_REVENUE - LO_SUPPLYCOST) AS profitFROM lineorder_flatWHERE (C_REGION = 'AMERICA') AND (S_REGION = 'AMERICA') AND ((P_MFGR = 'MFGR#1') OR (P_MFGR = 'MFGR#2'))GROUP BY     year,    C_NATIONORDER BY     year ASC,    C_NATION ASC
┌─year─┬─C_NATION──────┬────────profit─┐1992 │ ARGENTINA     │ 10419830420661992 │ BRAZIL        │ 1031193572794│ .... │ ......        │  ............ │1998 │ PERU          │  6039800448271998 │ UNITED STATES │  605069471323└──────┴───────────────┴───────────────┘
35 rows in set. Elapsed: 5.066 sec. Processed 600.04 million rows, 8.41 GB (118.43 million rows/s., 1.66 GB/s.)  

扫描行数:600,040,000 大约6亿

耗时(秒):5.066

查询列数:4

结果行数:35

Query 4.2

​​​​​​​​​​​​​​

SELECT     toYear(LO_ORDERDATE) AS year,    S_NATION,    P_CATEGORY,    sum(LO_REVENUE - LO_SUPPLYCOST) AS profitFROM lineorder_flatWHERE (C_REGION = 'AMERICA') AND (S_REGION = 'AMERICA') AND ((year = 1997) OR (year = 1998)) AND ((P_MFGR = 'MFGR#1') OR (P_MFGR = 'MFGR#2'))GROUP BY     year,    S_NATION,    P_CATEGORYORDER BY     year ASC,    S_NATION ASC,    P_CATEGORY ASC
┌─year─┬─S_NATION──────┬─P_CATEGORY─┬───────profit─┐1997 │ ARGENTINA     │ MFGR#111023699502151997 │ ARGENTINA     │ MFGR#12103052774082│ .... │ .........     │ .......    │ ............ │1998 │ UNITED STATES │ MFGR#24607793883451998 │ UNITED STATES │ MFGR#2560042710566└──────┴───────────────┴────────────┴──────────────┘
100 rows in set. Elapsed: 0.826 sec. Processed 144.42 million rows, 2.17 GB (174.78 million rows/s., 2.63 GB/s.)

扫描行数:144,420,000 大约1亿4千多万

耗时(秒):0.826

查询列数:4

结果行数:100

性能测试结果汇总

在当前软硬件环境下,扫描 6 亿多行数据,常见的分析语句首次运行最慢在 8 秒左右能返回结果,相同的分析逻辑更换条件再次查询的时候效率有明显的提升,可以缩短到 1 秒左右,如果只是简单的列查询没有加减乘除、聚合等逻辑,扫描全表 6 亿多行数据首次查询基本可以在 2 秒内执行完成。

声明:本文内容由网友自发贡献,不代表【wpsshop博客】立场,版权归原作者所有,本站不承担相应法律责任。如您发现有侵权的内容,请联系我们。转载请注明出处:https://www.wpsshop.cn/w/我家自动化/article/detail/702742
推荐阅读
相关标签
  

闽ICP备14008679号