R 根据矩阵中的单元格值删除列，并消除下标越界错误

我有一个问题，我必须删除Excel文件中的某些列，这取决于使用R的某些单元格值

我导入了Excel文件

data = as.matrix(read.csv('data.csv', header = FALSE))

矩阵的维数为：

dim（数据）=106474*81

我需要做什么

• 如果任何单元格中的值

  [行=1，列=i]='a'->

  删除列
• 如果任何单元格中的值

  [行=1，列=i]='B'->

  删除列
• 如果任何单元格中的值

  [行=2，列=i]>30->

  删除列

到目前为止我做了什么

for (i in 1:ncol(data)) { 
  if (data[1,i]=='A') {
    data =  data [,-i]
  } else if (data[1,i]=='B') {
    data =  data [,-i]
  } else if (data[2,i] > 30) {
    data =  data [,-i]
  } 
}

我遇到的问题如下：

当我把矩阵读给r时，它把所有的东西都读成字符？如何将数据[2，i]行转换为 数字的。我提供了如何将其转换为数字。但是有更好的办法吗

row2 = as.numeric(as.matrix(fly_data[2,])) 
fly_data[2,] <- row2

它总是给人一种感觉

数据[1，i]中出错：下标超出范围

---

也许你可以试试下面的代码

r <- subset(data, select = -which(as.character(unlist(data[1,]))%in% c("A","B") | as.numeric(as.character(unlist(data[2,])))>30))

数据

data <- structure(list(X1 = structure(2:1, .Names = c("r1", "r2"), .Label = c("7", 
"A"), class = "factor"), X2 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("6", "D"), class = "factor"), X3 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("24", "C"), class = "factor"), X4 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("13", "A"), class = "factor"), X5 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("27", "B"), class = "factor"), X6 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("17", "A"), class = "factor"), X7 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("25", "C"), class = "factor"), X8 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("35", "C"), class = "factor"), X9 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("13", "B"), class = "factor"), X10 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("27", "B"), class = "factor")), class = "data.frame", row.names = c("r1", 
"r2"))

> data
   X1 X2 X3 X4 X5 X6 X7 X8 X9 X10
r1  A  D  C  A  B  A  C  C  B   B
r2  7  6 24 13 27 17 25 35 13  27

数据
x1x2x3x4x5x6x7x8x9x10
r1 A D C A B A C B B
r2 7 6 24 13 27 17 35 13 27

当您阅读时，您可能会得到更多/更好的答案。您试图阅读的文件是

.csv

？那么我就不会称之为“excel”文件，这只会导致混乱。至于为什么所有的值都是字符的问题：一个矩阵只能包含一种类型的值，所以如果其中任何一个值是字符串，那么整个矩阵将被转换为字符。对混合类型使用

data.frame

或

data.table:：data.table

。然后指定for循环从1变为

ncol（data）

，但是在循环中从

数据中删除列

——这就是为什么

下标超出范围

错误的原因。你可以先复制一份，然后对它进行变异，而不是原来的。非常感谢。成功了。1.通过使用jogo as.numeric（数据[2，i]）>30 2解决了数值问题。通过使用数据的副本运行循环来解决越界问题。因此，原始矩阵运行for循环，而新矩阵正按照@dario的建议进行变异。

i30）；数据新
> r 
   X2 X3 X7
r1  D  C  C
r2  6 24 25

data <- structure(list(X1 = structure(2:1, .Names = c("r1", "r2"), .Label = c("7", 
"A"), class = "factor"), X2 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("6", "D"), class = "factor"), X3 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("24", "C"), class = "factor"), X4 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("13", "A"), class = "factor"), X5 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("27", "B"), class = "factor"), X6 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("17", "A"), class = "factor"), X7 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("25", "C"), class = "factor"), X8 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("35", "C"), class = "factor"), X9 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("13", "B"), class = "factor"), X10 = structure(2:1, .Names = c("r1", 
"r2"), .Label = c("27", "B"), class = "factor")), class = "data.frame", row.names = c("r1", 
"r2"))

> data
   X1 X2 X3 X4 X5 X6 X7 X8 X9 X10
r1  A  D  C  A  B  A  C  C  B   B
r2  7  6 24 13 27 17 25 35 13  27