Question

我有数据框DF：

DF <- data.frame(V1 = factor(c("Yes", "No", "Yes", "No", "No")),
                 V2 = factor(c("Yes", "No", "No", "Yes", "No")),
                 Location = factor(c("London", "Paris", "No", "Dallas", "No")),
                 V3 = factor(c("No", "Yes", "No", "No", "No")),
                 V4 = factor(c("No", "Yes", "No", "No", "No")))

我想将变量"No"，V1，V2和V3（而不是V4）中的值Location更改为"X"。我可以轻松地在每一列中手动更改级别的名称，但这在大型数据集中非常耗时。但是，如果我使用revalue，那么每个我想保持不变的"No"，包括Location中的那些，都将更改为"X"：

library("plyr")
as.data.frame(lapply(DF, function(x) { revalue(x, c("No"="X")) }))

是否有一种方法可以根据变量在数据集中的位置（此处为列1：2和4：5）来指定变量，以对其进行重命名？

Answer 1

另一种使用dplyrs quosure style lambda ~ fun(.)作为.funs参数并结合forcats :: fct_recode：

的解决方案

library("dplyr")
library("forcats")

(DF <- DF %>%
    mutate_at(vars(-Location), ~fct_recode(., "X" = "No")))

#    V1  V2 Location  V3  V4
# 1 Yes Yes   London   X   X
# 2   X   X    Paris Yes Yes
# 3 Yes   X       No   X   X
# 4   X Yes   Dallas   X   X
# 5   X   X       No   X   X

dplyr 1.0的更新：

新的across()取代了mutate_at之类的“作用域变体”。 across()可以轻松地将相同的转换应用于多个列，从而允许您在summarise（）和mutate（）中使用select（）语义

此处适用于问题，以下是实现此目的的两种变体：

DF %>%
  mutate(across((!Location), ~fct_recode(., "X" = "No")))

DF %>%
  mutate(across(c(1:2,4:5), ~fct_recode(., "X" = "No")))

Answer 2

也许有人可以提出一个更优雅的解决方案，但是以下一种可行的解决方案（无需手动更改每个变量）如下：

change.vec = c("V1", "V2", "V3", "V4")

for(i in 1:length(change.vec)) {
    levels(DF[,change.vec[i]]) = c("X", "Yes")  
}

>DF
 V1  V2 Location  V3  V4
 1 Yes Yes   London   X   X
 2   X   X    Paris Yes Yes
 3 Yes   X       No   X   X
 4   X Yes   Dallas   X   X
 5   X   X       No   X   X

Answer 3

只需指定要将revalue函数应用于的列号：

cols_to_update <- c(1:2,4:5)
DF[, cols_to_update] <- lapply(DF[,cols_to_update], function(x) plyr::revalue(x, c("No"="X")))

Answer 4

您也可以使用tidyverse方法进行此操作：

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(plyr)
#> -------------------------------------------------------------------------
#> You have loaded plyr after dplyr - this is likely to cause problems.
#> If you need functions from both plyr and dplyr, please load plyr first, then dplyr:
#> library(plyr); library(dplyr)
#> -------------------------------------------------------------------------
#> 
#> Attaching package: 'plyr'
#> The following objects are masked from 'package:dplyr':
#> 
#>     arrange, count, desc, failwith, id, mutate, rename, summarise,
#>     summarize
DF <- data.frame(V1 = factor(c("Yes", "No", "Yes", "No", "No")),
                 V2 = factor(c("Yes", "No", "No", "Yes", "No")),
                 Location = factor(c("London", "Paris", "No", "Dallas", "No")),
                 V3 = factor(c("No", "Yes", "No", "No", "No")),
                 V4 = factor(c("No", "Yes", "No", "No", "No")))
(DF <- DF %>%
    mutate_at(.vars = vars(-Location),
              .funs = function(t) revalue(x = t,
                                          replace = c("No" = "X"))))
#>    V1  V2 Location  V3  V4
#> 1 Yes Yes   London   X   X
#> 2   X   X    Paris Yes Yes
#> 3 Yes   X       No   X   X
#> 4   X Yes   Dallas   X   X
#> 5   X   X       No   X   X

^{由reprex package（v0.2.1）于2019-03-17创建}

重命名数据框中多个特定因素中的级别

4 个答案: