如何将integer64值的数据帧转换为矩阵? [英] How to convert a data frame of integer64 values to be a matrix?

问题描述

我有一个数据框架，完全由 integer64 我想转换为矩阵的列.

I have a data frame consisting entirely of integer64 columns that I'd like to convert to be a matrix.

library(bit64)
(dfr <- data.frame(x = as.integer64(10^(9:18))))
##                      x
## 1           1000000000
## 2          10000000000
## 3         100000000000
## 4        1000000000000
## 5       10000000000000
## 6      100000000000000
## 7     1000000000000000
## 8    10000000000000000
## 9   100000000000000000
## 10 1000000000000000000

很遗憾， as.matrix 没有给出正确答案.

Unfortunately, as.matrix doesn't give the correct answer.

(m <- as.matrix(dfr))
##                   x
##  [1,] 4.940656e-315
##  [2,] 4.940656e-314
##  [3,] 4.940656e-313
##  [4,] 4.940656e-312
##  [5,] 4.940656e-311
##  [6,] 4.940656e-310
##  [7,] 4.940656e-309
##  [8,] 5.431165e-308
##  [9,] 5.620396e-302
## [10,] 7.832953e-242

问题似乎是integer64值存储为具有"integer64"类属性的数字值(加上一些魔术使它们能够正确打印和执行算术)，并被as.matrix剥离.

The problem seems to be that integer64 values are stored as numeric values with an "integer64" class attribute (plus some magic to make them print and do arithmetic correctly) that gets stripped by as.matrix.

我不能只做class(m) <- "integer64"，因为那会改变矩阵对象的类而不是其内容.

I can't just do class(m) <- "integer64" because that changes the class of the matrix object not its contents.

同样，mode(m) <- "integer64"给出错误的答案，并且typeof(m) <- "integer64"和storage.mode(m) <- "integer64"抛出错误.

Likewise, mode(m) <- "integer64" gives the wrong answer and typeof(m) <- "integer64" and storage.mode(m) <- "integer64" throw errors.

我当然可以通过将列转换为double(dfr$x <- as.double(dfr$x))来解决问题，但是感觉应该有一种适当的方法来实现此目的.

Of course I could just circumvent the problem by converting the columns to double (dfr$x <- as.double(dfr$x)) but it feels like there ought to be a way to do this properly.

如何获取integer64值的矩阵?

推荐答案

对于原始向量，直接分配dim属性似乎可行:

For a raw vector, assigning the dim attribute directly seems to work:

> z <- as.integer64(1:10)
> z
integer64
 [1] 1  2  3  4  5  6  7  8  9  10
> dim(z) <- c(10, 1)
> z
integer64
      [,1]
 [1,] 1   
 [2,] 2   
 [3,] 3   
 [4,] 4   
 [5,] 5   
 [6,] 6   
 [7,] 7   
 [8,] 8   
 [9,] 9   
[10,] 10  

对于数据框，cbind对列也起作用:

For a data frame, cbinding the columns also works:

> df <- data.frame(x=as.integer64(1:5), y=as.integer64(6:10))
> df
  x  y
1 1  6
2 2  7
3 3  8
4 4  9
5 5 10
> cbind(df$x, df$y)
integer64
     [,1] [,2]
[1,] 1    6   
[2,] 2    7   
[3,] 3    8   
[4,] 4    9   
[5,] 5    10  

因此，对于任意数量的列，do.call是解决方法:

So, for an arbitrary number of columns, do.call is the way to go:

> do.call(cbind, df)
integer64
     x y 
[1,] 1 6 
[2,] 2 7 
[3,] 3 8 
[4,] 4 9 
[5,] 5 10

这篇关于如何将integer64值的数据帧转换为矩阵?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！