将矩阵转换为r中的数据帧

Question

我正在尝试将矩阵转换为数据帧,并使用数据框中的变量在矩阵中使用列名和行名.
这是样本

sample = matrix(c(1,NA,NA,2,NA,3,NA,NA,5,NA,NA,6,NA,NA,NA,NA,8,NA,3,1),ncol = 4)
colnames(sample) = letters[1:4]
row.names(sample) = letters[22:26]

我的数据集有很多,NA所以我试图删除NA数据框中的所有内容.

所以这是我渴望的输出,

 data.frame(col = c("v","v","w","w","y","y","y","z"),
           row = c("a","b","c","c","a","b","d","d"), 
           value = c(1,3,6,8,2,5,3,1))

Answer 1

使用meltreshape2包重新整形,然后清除NA.最后,做一些格式化的东西来获得你想要的输出(订购,设置colnames ......).

> library(reshape2)
> df <- na.omit(melt(sample))  # reshaping
> df <- df[order(df$Var1), ]   # ordering
> colnames(df) <- c("col", "row", "value") # setting colnames
> df                           # getting desired output
   col row value
1    v   a     1
6    v   b     3
12   w   c     6
17   w   d     8
4    y   a     2
9    y   b     5
19   y   d     3
20   z   d     1

用dplyr和magrittr

> library(magrittr)
> library(dplyr)
> sample %>%  melt %>% 
    na.omit %>% 
    arrange(., Var1) %>%
    setNames(c('col', 'row', 'value'))

 col row value
1   v   a     1
2   v   b     3
3   w   c     6
4   w   d     8
5   y   a     2
6   y   b     5
7   y   d     3
8   z   d     1