使用变量名作为函数参数

Question

我想编写一个独特的函数来（除其他外）在不同的阈值下为不同的变量过滤数据库。我找到了一种方法来指示我想在函数本身中过滤的变量，但不确定是不是最好的方法。

我该怎么办？

example_db <- data.frame(name=c("A","B","C"), 
                         value_1=c(1,2,3), 
                         value_2=c(2,3,1))

advanced_filter <- function(data,variable,limit){
  
  require(dplyr)
  data <- data %>% 
    dplyr::filter(variabe>limit) 
  
  return(data)
}

预期结果：

advanced_filter(example_db,value_1,2)

name value_1 value_2
1    C       3       1

我的尝试：

advance_filter <- function(data,variable,limit){
  require(dplyr)
  f <- paste(variable, ">",  limit)
  data <- data %>% 
    dplyr::filter_(f) 
  
  return(data)
}


advance_filter(example_db,"value_1",2)

Answer 1

也许你把它弄得比需要的更复杂了：

advanced_filter <- function(data, variable, limit) {
  data[data[variable] > limit,]
}

advanced_filter(example_db, "value_1", 2)
#>   name value_1 value_2
#> 3    C       3       1

^{由 reprex package (v2.0.1)}

于 2022-01-28 创建

Answer 2

Allan Cameron 的答案显然是正确的，只需要基础 R，为了后代，这里是整洁的版本。

example_db <- data.frame(name=c("A","B","C"), 
                         value_1=c(1,2,3), 
                         value_2=c(2,3,1))



advanced_filter <- function(data,variable,limit){
  require(dplyr)
  vbl <- enquo(variable)
  data %>% 
    dplyr::filter(!!vbl > limit) 
}

advanced_filter(example_db,value_1,2)
#> Loading required package: 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
#>   name value_1 value_2
#> 1    C       3       1

^{由 reprex package (v2.0.1)}

于 2022-01-28 创建

或者，按照下面@TimTeaFan 的评论：

example_db <- data.frame(name=c("A","B","C"), 
                         value_1=c(1,2,3), 
                         value_2=c(2,3,1))



advanced_filter <- function(data,variable,limit){
  require(dplyr)
  data %>% 
    dplyr::filter({{variable}} > limit) 
}

advanced_filter(example_db,value_1,2)
#> Loading required package: 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
#>   name value_1 value_2
#> 1    C       3       1

^{由 reprex package (v2.0.1)}

于 2022-01-28 创建

Answer 3

如果您希望能够将矢量名称用作字符串，无论是完全限定的 - 还是仅作为符号，您可以使用以下 Base R 解决方案：

# Function:
advanced_filter <- function(data, variable, limit) {
  if(is.character(substitute(variable))){
    variable_str <- variable
  }else{
    variable_str <- gsub(".*\$", "", deparse(substitute(variable)))
  }
  res <- data[data[,variable_str] > limit,]
  return(res)
}
# Application:
advanced_filter(example_db, "value_1", 2)
advanced_filter(example_db, value_1, 2)
advanced_filter(example_db, example_db$value_1, 2)

使用变量名作为函数参数

Use a variable name as function argument

variables

r

filter

dplyr

tidyverse