Question

我有两个矩阵m1和m2具有相同的结构，现在我需要创建2个新的矩阵
1）第一个（称为Partenza）具有m1所有与m2不同的行 2）第二个（calle Arrivo）具有来自m2的相同行

m1 
    row_num  datoA        datoB
    1          p           f
    3          h           b
    5          c           m
    6          c           r
    9          m           f
    14         a           b

m2 
    row_num  datoA        datoB
    1          p           f
    3          h           b
    5          c           g
    6          a           r
    9          m           f
    14         x           j

我的结果应该是：

Partenza (taken from m1)
    row_num  datoA        datoB
    5          c           m
    6          c           r
    14         a           b

Arrivo (taken from m2)
    row_num  datoA        datoB
    5          c           g
    6          a           r
    14         x           j

我试过

zzz <- setdiff(m1,m2)  
partenza<-m1[m1[,"ROW_NUM"] %in% zzz,]
arrivo<-  m1[m1[,"ROW_NUM"] %in% zzz,]

但它无法告诉我zzz总是空的（我相信它应该不是！）

Answer 1

使用dplyr包中的anti_join功能：

package(dplyr)

m1 <- data.frame(
  row_num = c(1,3,5,6,9,14),
  datoA = c("p","h","c","c","m","a"),
  datoB = c("f","b","m","r","f","b")
  )

m2 <- data.frame(
  row_num = c(1,3,5,6,9,14),
  datoA = c("p","h","c","a","m","x"),
  datoB = c("f","b","g","r","f","j")
)

Partenza <- anti_join(m1,m2) %>% arrange(row_num)
Arrivo <- anti_join(m2,m1) %>% arrange(row_num)

Answer 2

在base R（不需要包裹）中，您可以执行以下操作：

diffrow<-sapply(1:nrow(m1),function(x) !all(m1[x,]==m2[x,]))
partenza<-m1[diffrow,]
arrivo<-m2[diffrow,]

另一个选项，使用包setdiff中的dplyr函数：

library(dplyr)
partenza<-setdiff(m1,m2)
arrivo<-setdiff(m2,m1)

在这两种情况下，你都会得到：

> partenza
#  row_num datoA datoB
#3       5     c     m
#4       6     c     r
#6      14     a     b
> arrivo
#  row_num datoA datoB
#3       5     c     g
#4       6     a     r
#6      14     x     j

数据：

m1<-structure(list(row_num = c(1L, 3L, 5L, 6L, 9L, 14L), datoA = c("p", "h", "c", "c", "m", "a"), datoB = c("f", "b", "m", "r", "f", "b")), .Names = c("row_num", "datoA", "datoB"), class = "data.frame", row.names = c(NA, -6L)) m2<-structure(list(row_num = c(1L, 3L, 5L, 6L, 9L, 14L), datoA = c("p", "h", "c", "a", "m", "x"), datoB = c("f", "b", "g", "r", "f", "j")), .Names = c("row_num", "datoA", "datoB"), class = "data.frame", row.names = c(NA, -6L))

R从矩阵中提取具有不同值的行

2 个答案: