Changing the levels of factor to numeric values:

Question

I have a vector with three levels, which are "M", "S", and "unknown"

I want to change them to numeric (integer) values like this:

1. If "M" then value 2 (integer)

2. if "S" then value 1 (integer)

3. if "unknown" then 0 (integer)

What I have tried

I was going to change them one by one like this:

cards$MaritalStatus[cards$MartialStatus == "M"] <- 2

However, I got exception:

Warning message:
In `[<-.factor`(`*tmp*`, cards$MartialStatus == "M", value = c(1L,  :
  invalid factor level, NA generated

Any help please?

Answer 1

我会用匹配来做到这一点：

match(cards$MaritalStatus, c("unknown", "S", "M")) - 1

Answer 2

Try

v1 <- factor(c("M","S","unknown","M","S"))
keyval <- setNames(0:2, c('unknown', 'S', 'M'))
as.numeric(keyval[as.character(v1)])
#[1] 2 1 0 2 1

Or

as.numeric(as.character(factor(v1, levels=c('M', 'S', 'unknown'),
           labels=c(2,1,0)) ))
#[1] 2 1 0 2 1

Answer 3

如果你有超过3个值，我不会采用这种方法但在这种情况下你可以使用嵌套的ifelse语句，如下所示：

ifelse(cards$MaritalStatus=="unknown", 0, ifelse(cards$MaritalStatus=="M", 2, 1))

Answer 4

Here is a more general approach (and maybe easier to understand):

1st - declare values and what they should be changed to.

namesList <- list(M=2, A=1, unknown=0)

2nd - apply changes

unlist(namesList[as.character(cards$MaritalStatus)])