r语言 - pmap 中的错误 UseMethod( "filter_") 中的错误：没有适用于类 "character" 对象的"filter_"方法。

Question

我编写了一个函数，当我输入输入变量作为数据框时，该函数工作正常。但是当我想使用 pmap 将输入作为数据框列表输入时，出现以下错误：

Error in UseMethod("filter_") : no applicable method for 'filter_' applied to an object of class "character"

这是导致错误的函数的数据和第一部分，我在函数的某些部分中使用了 y 和参数，此处未显示：

x <- tibble::tibble(x1 = sample(0:1, 8, replace = TRUE),
x2 = sample(0:25, 8, replace = FALSE),
x3 = sample(1:3, 8, replace = TRUE),
strata =c("a", "b", "c", "d", "a", "b", "c", "d"))
y <- tibble::tibble(rate = sample(0:1, 8, replace = TRUE),
strata =c("a", "b", "c", "d", "a", "b", "c", "d") )
a <- tibble::tibble(sample(10:80, 4, replace = FALSE))
example <- function(x, y, a , d){
CR <- x %>% filter(x1, x2>0) %>%
group_by(x3) %>%
summarise(avg_revenue = mean(x2), revenue = sum(x2))
return(CR)
}
example(x,y,a, d = 0.1)

但是当我在这个函数上调用 pmap 时：

df <- tibble::tibble(x = x %>% group_by(strata) %>% nest(),
y = y %>% group_by(strata) %>% nest(),
a = a)
pmap(df, example, d= 0.1)

我收到上面提到的错误。

Answer 1

我不相信df正在创建您希望它创建的df。我相信这能做到你想要的...如果我正确理解了这个问题。但是y在您的函数中没有使用，所以我不清楚它的目的是什么。我相信还有一种更好的方法可以使用map和nest来做到这一点，但我再次不确定您要做什么。

library(tidyverse)
x <- tibble::tibble(x1 = sample(0:1, 8, replace = TRUE),
x2 = sample(0:25, 8, replace = FALSE),
x3 = sample(1:3, 8, replace = TRUE),
strata =c("a", "b", "c", "d", "a", "b", "c", "d"))
y <- tibble::tibble(rate = sample(0:1, 8, replace = TRUE),
strata =c("a", "b", "c", "d", "a", "b", "c", "d") )
a <- tibble::tibble(a = sample(10:80, 4, replace = FALSE))
example <- function(x, y, a , d){
CR <- x %>% filter(x1, x2>0) %>%
group_by(x3) %>%
summarise(avg_revenue = mean(x2), revenue = sum(x2))
return(CR)
}
example(x,y,a, d = 0.1)
#> # A tibble: 1 x 3
#>      x3 avg_revenue revenue
#>   <int>       <dbl>   <int>
#> 1     1           5      10
df <- bind_cols(x, select(y, rate)) %>% 
group_by(strata) %>% 
nest(x = c(x1, x2, x3), 
y = c(rate)) %>% 
bind_cols(a) %>% ungroup()
pmap(select(df, -strata), example)
#> [[1]]
#> # A tibble: 0 x 3
#> # … with 3 variables: x3 <int>, avg_revenue <dbl>, revenue <int>
#> 
#> [[2]]
#> # A tibble: 0 x 3
#> # … with 3 variables: x3 <int>, avg_revenue <dbl>, revenue <int>
#> 
#> [[3]]
#> # A tibble: 1 x 3
#>      x3 avg_revenue revenue
#>   <int>       <dbl>   <int>
#> 1     1           4       4
#> 
#> [[4]]
#> # A tibble: 1 x 3
#>      x3 avg_revenue revenue
#>   <int>       <dbl>   <int>
#> 1     1           6       6
pmap_dfr(select(df, -strata), example, d = 0.1, .id = 'strata')
#> # A tibble: 2 x 4
#>   strata    x3 avg_revenue revenue
#>   <chr>  <int>       <dbl>   <int>
#> 1 3          1           4       4
#> 2 4          1           6       6

^{创建于 2019-12-17 由 reprex 软件包 (v0.3.0(}

Answer 2

正如 CLedbetter 在他们的有用答案中也提到的那样，当要pmap的输入数据帧df的格式不正确时，就会出现此错误。pmap期望df仅具有它所操作的函数已知的列。 为此，我用inner_join编辑了df，然后我们仍然有函数example()不知道的列strata。 正如在R中pmap函数的帮助中提到的，为了使pmap函数忽略函数example()不使用的列， 我用了"..."在example()的定义中，以便 pmap 可以跳过函数中未使用的数据帧的第一列strata。

因此，更新后的代码将是：

x <- tibble::tibble(x1 = sample(0:1, 8, replace = TRUE),
x2 = sample(0:25, 8, replace = FALSE),
x3 = sample(1:3, 8, replace = TRUE),
strata =c("a", "b", "c", "d", "a", "b", "c", "d"))
y <- tibble::tibble(rate = sample(0:1, 8, replace = TRUE),
strata =c("a", "b", "c", "d", "a", "b", "c", "d") )
a <- tibble::tibble(sample(10:80, 4, replace = FALSE))
# Note the addition of the "..." to the function input definition
example <- function(x, y, a , d, ...){
CR <- x %>% filter(x1, x2>0) %>%
group_by(x3) %>%
summarise(avg_revenue = mean(x2), revenue = sum(x2))
return(CR)
}
example(x,y,a, d = 0.1)
# Note the change in the reformatting of df with an inner_join
df <- inner_join(x %>% group_by(strata) %>% nest(),
y %>% group_by(strata) %>% nest(), 
by = "strata") %>% rename(x = data.x, y = data.y )
# with these changes pmap produces the output 
pmap(df, example, d= 0.1)