R语言列操作函数

目录

一.dplyr包

1.新增变量和变量重新赋值

2.筛选行

3.筛选列

4.分组计算

5.管道操作符

6.连接数据框

二.tidyr

1.列的分裂

2.列的合并

3.宽数据转长数据

4.长数据转宽数据

---

一.dplyr包

1.新增变量和变量重新赋值

> head(ToothGrowth)
   len supp dose
1  4.2   VC  0.5
2 11.5   VC  0.5
3  7.3   VC  0.5
4  5.8   VC  0.5
5  6.4   VC  0.5
6 10.0   VC  0.5
> ToothGrowth2<-mutate(ToothGrowth,len=len^2,nv=1:nrow(ToothGrowth),
nv2=ifelse(nv>median(nv),"H","L"))

> head(ToothGrowth2)
     len supp dose nv nv2
1  17.64   VC  0.5  1   L
2 132.25   VC  0.5  2   L
3  53.29   VC  0.5  3   L
4  33.64   VC  0.5  4   L
5  40.96   VC  0.5  5   L
6 100.00   VC  0.5  6   L

2.筛选行

> ToothGrowth3<-filter(ToothGrowth2,nv %in% 1:50,nv2=="H")

> head(ToothGrowth3)
     len supp dose nv nv2
1 231.04   OJ  0.5 31   H
2 462.25   OJ  0.5 32   H
3 309.76   OJ  0.5 33   H
4  94.09   OJ  0.5 34   H
5 210.25   OJ  0.5 35   H
6 100.00   OJ  0.5 36   H

3.筛选列

#只显示第二列和第四列的数据
ToothGrowth4 <- select(ToothGrowth3,c(2,4))

4.分组计算

> summarise(ToothGrowth,len_max=max(len))
  len_max
1    33.9
#会按照supp进行分组
> summarise(group_by(ToothGrowth,supp),len_max=max(len))
# A tibble: 2 × 2
  supp  len_max
     
1 OJ       30.9
2 VC       33.9
> summarise(group_by(ToothGrowth,dose),len_max=max(len))
# A tibble: 3 × 2
   dose len_max
     
1   0.5    21.5
2   1      27.3
3   2      33.9
> summarise(group_by(ToothGrowth,dose,supp),len_max=max(len))
`summarise()` has grouped output by 'dose'. You can override using the `.groups`
argument.
# A tibble: 6 × 3
# Groups:   dose [3]
   dose supp  len_max
      
1   0.5 OJ       21.5
2   0.5 VC       11.5
3   1   OJ       27.3
4   1   VC       22.5
5   2   OJ       30.9
6   2   VC       33.9

5.管道操作符

> library(magrittr)
#将左边的参数作为右边参数的第一个参数
> ToothGrowth %>%
+ mutate(nv=1:nrow(ToothGrowth))%>%
+ filter(nv %in% 1:50)%>%
+ select(1:2) %>%
+ group_by(supp)%>%
+ summarise(len_max=max(len))%>%
+ as.data.frame()

6.连接数据框

> df1<-data.frame(c1=2:5,c2=LETTERS[2:5])
> df1
  c1 c2
1  2  B
2  3  C
3  4  D
4  5  E
> df2<-data.frame(c3=LETTERS[c(2:3,20:23)],c4=sample(1:100,size=6))
> df2
  c3 c4
1  B 62
2  C 20
3  T 29
4  U 42
5  V 60
6  W 65
#left_join(df1,df2,by=c('c2'=='c3')):df1左连接于df2 
> df1 %>% left_join(df2,by=c('c2'='c3'))
  c1 c2 c4
1  2  B 62
2  3  C 20
3  4  D NA
4  5  E NA
> df1 %>% right_join(df2,by=c('c2'='c3'))
  c1 c2 c4
1  2  B 62
2  3  C 20
3 NA  T 29
4 NA  U 42
5 NA  V 60
6 NA  W 65
> df1 %>% full_join(df2,by=c('c2'='c3'))
  c1 c2 c4
1  2  B 62
2  3  C 20
3  4  D NA
4  5  E NA
5 NA  T 29
6 NA  U 42
7 NA  V 60
8 NA  W 65
> df1 %>% inner_join(df2,by=c('c2'='c3'))
  c1 c2 c4
1  2  B 62
2  3  C 20

二.tidyr

1.列的分裂

> library(tidyr)
> df3<-data.frame(c5=paste(letters[1:3],1:3,sep="-"),
+ c6=paste(letters[1:3],1:3,sep="."),
+ c4=c("B","B","B"),
+ c3=c("H","M","L"))
> df3
   c5  c6 c4 c3
1 a-1 a.1  B  H
2 b-2 b.2  B  M
3 c-3 c.3  B  L
> df4<-df3%>%
+ separate(col=c5,sep="-",into=c("c7","c8"),remove=F)%>%
+ separate(col=c6,sep="\\.",into=c("c9","c10"),remove=T)
#\\表示转义字符
> df4
   c5 c7 c8 c9 c10 c4 c3
1 a-1  a  1  a   1  B  H
2 b-2  b  2  b   2  B  M
3 c-3  c  3  c   3  B  L

2.列的合并

> df4%>%
#remove表示将原来操作的列保留下来
+ unite(col="c11",c("c7","c8"),sep="_",remove=F)%>%
+ unite(col="c12",c("c9","c10"),sep=".",remove=T)
   c5 c11 c7 c8 c12 c4 c3
1 a-1 a_1  a  1 a.1  B  H
2 b-2 b_2  b  2 b.2  B  M
3 c-3 c_3  c  3 c.3  B  L

个人以为以下两个函数特别重要，特别有用，特别是在绘制线性图的时候

宽数据：列数变多了，names_from，表示列名的来源，value_from(),表示列值来源

从长数据到宽数据为from

长数据：行数变多了，names_to，表示列名加入到行中的去向，value_from(),表示列值加入到行中的去向，从宽数据到长数据为to

3.宽数据转长数据

> set.seed(42)
> df5<-data.frame(time=rep(2011:2013,each=3),
+ area=rep(letters[1:3],times=3),
+ pop=sample(100:1000,9),
+ den=round(rnorm(9,mean=3,sd=0.1),2),
+ mj=sample(8:12,9,replace = T))
#replace=T,有放回的取
> df5

 time area pop  den mj
1 2011    a 660 2.99 12
2 2011    b 420 3.15 12
3 2011    c 252 2.99 11
4 2012    a 173 3.20  9
5 2012    b 327 2.99 11
6 2012    c 245 3.13 10
7 2013    a 733 3.23  9
8 2013    b 148 2.86  8
9 2013    c 227 2.97  9

> df6<-df5%>%
#将最后三列数据合并为一列
+ pivot_longer(cols=-c(1:2),
#列名称
+ names_to="varb",
#列值
+ values_to="value")
> df6
# A tibble: 27 × 4
    time area  varb   value
       
 1  2011 a     pop   660   
 2  2011 a     den     2.99
 3  2011 a     mj     12   
 4  2011 b     pop   420   
 5  2011 b     den     3.15
 6  2011 b     mj     12   
 7  2011 c     pop   252   
 8  2011 c     den     2.99
 9  2011 c     mj     11   
10  2012 a     pop   173   
# ℹ 17 more rows
# ℹ Use `print(n = ...)` to see more rows
> 

4.长数据转宽数据

> df6%>%
+ pivot_wider(names_from=c(area,varb),values_from = value)
# A tibble: 3 × 10
   time a_pop a_den  a_mj b_pop b_den  b_mj c_pop c_den  c_mj
           
1  2011   660  2.99    12   420  3.15    12   252  2.99    11
2  2012   173  3.2      9   327  2.99    11   245  3.13    10
3  2013   733  3.23     9   148  2.86     8   227  2.97     9
>