How can I replace values in one column with values from another out of several options, when the first column contains the name of the other column?

Say I have the table below. I want to replace the value in the column "nameselect" with the respective value from either "name_01" or "name_02". Which of these columns should be chosen is indicated by the original value in "nameselect". As a result, I want "nameselect" to contain just names - in this case, Ann and Claire. I'd also be ok with a solution creating a new column with the names. How do I best go about that?
In my actual data, there's more name columns, so ideally nothing that involves copy-pasting every possible column name. I should also mention that it should work even when there are NAs in "nameselect"

| nameselect | name_01        | name_02

| --------   | -------------- |-------------- |

| name_01    | Ann            |Bernie         |

| name_02    | Beth           |Claire         |

With dplyr this works, but not when there are NAs:

df %>% 
  rowwise() %>% 
  mutate(result = get(nameselect)) %>% 
  ungroup()

Any way I can adapt this solution to work despite NAs?

You may have multiple columns, in which case a one-liner general solution using only base R would be:

within(df, nameselect <- sapply(seq(nrow(df)), \(i) df[i, df$nameselect[i]]))
#>   nameselect name_01 name_02
#> 1        Ann     Ann  Bernie
#> 2     Claire    Beth  Claire

See if this works considering you only have two columns

df[,'nameselect']=ifelse(df[,'nameselect']=='name_01',df[,'name_01'],df[,'name_01'])

As Limey said, it's quite straightforward:

df %>% 
  mutate(
    nameselect = ifelse(nameselect == "name_01", name_01, name_02))

# or
df$nameselect <- ifelse(df$nameselect == "name_01", df$name_01, df$name_02)

# or 
df %>% 
  rowwise() %>% 
  mutate(nameselect = get(nameselect))

# or 
df$nameselect <- map_vec(seq_along(df$nameselect), ~ df[[df$nameselect[.x]]][.x])

Based on this solution to a similar issue, https://stackoverflow.com/questions/67678405/r-lookup-values-of-a-column-defined-by-another-columns-values-in-mutate works with case_when even with NAs. Thanks for the suggestion!

x <- switch_cols <- function(var) {
  
  vals <- unique(var)
  
  name <- deparse(substitute(var))
  
  formulae <- lapply(
    sprintf("%s == '%s' ~ %s", name, vals, vals), 
    as.formula, 
    env = parent.frame()
  )
  
  case_when(!!!formulae)
  
}

df %>% 
  mutate(result = switch_cols(nameselect))

</details>



# 答案5
**得分**: 0

这是一个向量化的基本R方法，适用于任意数量的列。

```R
df$result <- df[cbind(seq(nrow(df)), match(df$nameselect, names(df)))]

df
#  nameselect name_01 name_02 result
#1    name_01     Ann  Bernie    Ann
#2    name_02    Beth  Claire Claire

我们创建了一个由行/列对组成的矩阵，以使用cbind从数据框中进行子集选择，其中seq(nrow(df))给出了行号，match(df$nameselect, names(df))给出了要子集选择的列号。

数据

如果您提供一个可重现的格式，将更容易提供帮助。

df <- structure(list(nameselect = c("name_01", "name_02"), name_01 = c("Ann", "Beth"), name_02 = c("Bernie", "Claire")), row.names = c(NA, -2L), class = "data.frame")

df$result <- df[cbind(seq(nrow(df)), match(df$nameselect, names(df)))]

df
#  nameselect name_01 name_02 result
#1    name_01     Ann  Bernie    Ann
#2    name_02    Beth  Claire Claire

df <- structure(list(nameselect = c("name_01", "name_02"), name_01 = c("Ann", 
"Beth"), name_02 = c("Bernie", "Claire")), row.names = c(NA, 
-2L), class = "data.frame")