r - 带 dplyr 的条件 rowMeans

Question

我想使用 dplyr 将条件 rowMeans 应用于每组 ID。可重现的表格如下:

dat <- as.Date("2021/08/04")
len <- 4
seq(dat, by = "day", length.out = len)

input <- data.frame(
  date = c(seq(dat, by = "day", length.out = len) , seq(dat, by = "day", length.out = len)),
  user_id = c(rep("aa",4),rep("bb",4)),
  var1 = c(1:4),
  var2 = c(4,7,23,9,0,0,0,0),
  var3 = c(0,0,0,0,4,8,2,7) 
)

逻辑是:对于每个 ID，首先选择数字列且它们的总和为 =!0，然后计算选定列的 rowMean那个ID。

所需的输出表如下:

 output = data.frame(
  date = c(seq(dat, by = "day", length.out = len) , seq(dat, by = "day", length.out = len)),
  user_id = c(rep("aa",4),rep("bb",4)),
  var1 = c(1:4),
  var2 = c(4,7,23,9,0,0,0,0),
  var3 = c(0,0,0,0,4,8,2,7),
  rowAverage = as.numeric(c(rowMeans(input[1:4,3:4]), rowMeans(input[5:8,c(3,5)])))
)

我想到的行如下，但我遇到了一些错误:

output = input %>%
    dplyr::group_by(user_id) %>%
    dplyr::mutate(rowAverage = rowMeans(select_if(function(x) {(is.numeric(x)) && (sum(x)=!0)})))

你能帮我改正这个吗？谢谢。

Answer 1

你很亲近。您也只需将数据提供给 select_if()。为此使用 cur_data()。

input %>%
  group_by(user_id) %>%
  mutate(rowAverage = rowMeans(
    select_if(cur_data(), \(x) is.numeric(x) && sum(x) != 0)
  ))
#> # A tibble: 8 × 6
#> # Groups:   user_id [2]
#>   date       user_id  var1  var2  var3 rowAverage
#>   <date>     <chr>   <int> <dbl> <dbl>      <dbl>
#> 1 2021-08-04 aa          1     4     0        2.5
#> 2 2021-08-05 aa          2     7     0        4.5
#> 3 2021-08-06 aa          3    23     0       13  
#> 4 2021-08-07 aa          4     9     0        6.5
#> 5 2021-08-04 bb          1     0     4        2.5
#> 6 2021-08-05 bb          2     0     8        5  
#> 7 2021-08-06 bb          3     0     2        2.5
#> 8 2021-08-07 bb          4     0     7        5.5