1. Introduction to
dplyr dplyr is a data analysis package of R language, similar to pandas in python, which can do very convenient data processing and analysis operations on dataframe type data. At first I was also surprised by the weird name dplyr. I found one of the explanations-d stands for dataframe-plyr is a homophone for plier in English
dplyr, like most R packages, is functional programming, which is very different from Python object-oriented programming. The advantage is that beginners are more likely to accept this kind of functional thinking, which is similar to an assembly line. Each function is a workshop, and multiple workshops work together to complete a production (data analysis) task.
In dplyr, there is a pipe symbol %>%. The left side of the symbol represents the input of data, and the right side represents the downstream data processing link.
2. Install and import
the p_load function of the dplyr library pacman library includes
- install.packages(“dplyr”)
- library(dplyr)
is easier to use
pacman::p_load("dplyr")
**3. 读取数据**
#Set the working directory setwd("/Users/thunderhit/Desktop/dplyr_learn")
#Import csv data aapl <- read.csv('aapl.csv',
header=TRUE,
sep=',',
stringsAsFactors = FALSE) %>% as_tibble()
head( aapl )
A tibble: 6 × 6
![](https://s4.51cto.com/images/blog/202012/30/e8ec6453c58ef833f05be96b520b3f66.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
查看数据类型
class(aapl)
'tbl_df'
'tbl'
'data.frame'
查看数据的字段
colnames (aapl)
'Date'
'Open'
'High'
'Low'
'Close'
'Volume'
查看记录数、字段数
dim (aapl)
251
6
**4. dplyr常用函数**
**4.1 Arrange**
对appl数据按照字段Volume进行降序排序
arrange(aapl, -Volume)
A tibble: 6 × 6
![](https://s4.51cto.com/images/blog/202012/30/b85165f87b036f85be4f04d0b9440d7f.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
我们也可以用管道符 %>% ,两种写法得到的运行结果是一致的,可能用久了会觉得管道符 %>% 可读性更强,后面我们都会用 %>% 来写代码。
aapl %>% arrange(-Volume)
A tibble: 6 × 6
![](https://s4.51cto.com/images/blog/202012/30/3bd99180a44d5679af4f84db43b961cd.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
**4.2 Select**
选取 Date、Close和Volume三列
aapl %>% select(Date, Close, Volume)
![](https://s4.51cto.com/images/blog/202012/30/e8a9d430742058ef02e7d40c63ab0f76.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
只选取Date、Close和Volume三列,其实另外一种表达方式是“排除Open、High、Low,选择剩下的字段的数据”。
aapl %>% select(-c("Open", "High", "Low"))
![](https://s4.51cto.com/images/blog/202012/30/05de7bd25514a8ddab6357ca7322ef7c.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
**4.3 Filter**
按照筛选条件选择数据
#Select from the data the transaction data with appl stock price greater than $150
aapl %>% filter(Close>=150)
![](https://s4.51cto.com/images/blog/202012/30/35242c4f0e45d7329e7033c0f64ef322.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
从数据中选择appl - 股价大于150美元 且 收盘价大于开盘价 的交易数据
aapl %>% filter((Close>=150) & (Close>Open))
![](https://s4.51cto.com/images/blog/202012/30/2a75e95cf0cceed0d7fdee04930053bd.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
**4.4 Mutate**
将现有的字段经过计算后生成新字段。
#Define the result of the best price High minus the lowest price Low as maxDif, and take log aapl
%>% mutate(maxDif = High-Low,
log_maxDif=log(maxDif))
![](https://s4.51cto.com/images/blog/202012/30/1383ab8f2704ff082184be58e9a8cd09.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
得到记录的位置(行数)
aapl %>% mutate(n=row_number())
![](https://s4.51cto.com/images/blog/202012/30/a6c2415852c4c7d5111adfb062909406.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
**4.5 Group_By**
对资料进行分组,这里导入新的 数据集 weather
#Import csv data weather <- read.csv('weather.csv',
header=TRUE,
sep=',',
stringsAsFactors = FALSE) %>% as_tibble()
weather
![](https://s4.51cto.com/images/blog/202012/30/e67eb1810297ed7cc07633bcf4a01f49.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
按照城市分组
weather %>% group_by(city)
![](https://s4.51cto.com/images/blog/202012/30/d6f71e09428e5262c64f7f9bcab295be.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
In order to let everyone see the effect of grouping, let’s calculate the average temperature by city.
weather %>% group_by(city) %>% summarise(mean_temperature = mean(temperature))
`summarise()` ungrouping output (override with `.groups` argument)
![](https://s4.51cto.com/images/blog/202012/30/93a8b1d75bd558a4bd70515761c2bbd9.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)
weather %>% summarise(mean_temperature = mean(temperature))
![](https://s4.51cto.com/images/blog/202012/30/8e9afc255c68f598d06a104355e145f5.png?x-oss-process=image/watermark,size_16,text_QDUxQ1RP5Y2a5a6i,color_FFFFFF,t_100,g_se,x_10,y_10,shadow_90,type_ZmFuZ3poZW5naGVpdGk=)