在 R 中使用 lag() 函数时用原始值替换 NA

Question

我正在使用 dplyr 的 lag() 函数，我试图找出不让 NA（而是取原始值）作为空白滞后单元格的默认值。

这是我的代码：

df <- data_frame(d1 = runif(10, 1, 5), 
                 d2 = runif(10, 2, 6),
                 d3 = runif(10, 3, 7),
                 d4 = runif(10, 4, 8),
                 d5 = runif(10, 5, 9),
                 d6 = runif(10, 6, 10),
                 d7 = runif(10, 7, 11),
                 d8 = runif(10, 8, 12)) %>% rownames_to_column() 
df %>%
  gather(key = "col", value = "val", -"rowname") %>%
  group_by(col) %>%
  mutate(new_col = ifelse(val >= lag(val, 2) + lag(val, 2)*0.4, NA, val))

如果我执行此代码（老实说，我很期待）它不起作用：

df %>%
      gather(key = "col", value = "val", -"rowname") %>%
      group_by(col) %>%
      mutate(new_col = if_else(val >= lag(val, 2, default = val) + lag(val, 2, default = val)*0.4, NA, val))

我缺少什么才能得到这个结果？

   rowname col     val new_col
   <chr>   <chr> <dbl>   <dbl>
 1 1       d1     1.31   **1.31**   
 2 2       d1     4.10   **4.10**   
 3 3       d1     3.81   NA   
 4 4       d1     4.52    4.52
 5 5       d1     3.89    3.89
 6 6       d1     1.01    1.01
 7 7       d1     2.68    2.68
 8 8       d1     2.81   NA   
 9 9       d1     1.18    1.18
10 10      d1     1.19    1.19
# ... with 70 more rows

感谢任何帮助！

Answer 1

您可以 replace n 滞后值与原始值。

library(dplyr)
n <- 2

df %>%
 tidyr::pivot_longer(cols = -rowname, values_to = 'val', names_to = 'col') %>%
 group_by(col) %>%
 mutate(new_col = if_else(val >= lag(val, n) + lag(val, n)*0.4, NA_real_, val),
        new_col  = replace(new_col, 1:n, val[1:n]))

Answer 2

coalesce就是针对这类问题而生的

library(tidyverse)

set.seed(42)

df <- data_frame(d1 = runif(10, 1, 5), 
                 d2 = runif(10, 2, 6),
                 d3 = runif(10, 3, 7),
                 d4 = runif(10, 4, 8),
                 d5 = runif(10, 5, 9),
                 d6 = runif(10, 6, 10),
                 d7 = runif(10, 7, 11),
                 d8 = runif(10, 8, 12)) %>% rownames_to_column() 
#> Warning: `data_frame()` is deprecated as of tibble 1.1.0.
#> Please use `tibble()` instead.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_warnings()` to see where this warning was generated.
df %>%
  gather(key = "col", value = "val", -"rowname") %>%
  group_by(col) %>%
  mutate(new_col = ifelse(val >= lag(val, 2) + lag(val, 2)*0.4, NA, val),
         new_col_no_na = coalesce(new_col,val))
#> # A tibble: 80 x 5
#> # Groups:   col [8]
#>    rowname col     val new_col new_col_no_na
#>    <chr>   <chr> <dbl>   <dbl>         <dbl>
#>  1 1       d1     4.66   NA             4.66
#>  2 2       d1     4.75   NA             4.75
#>  3 3       d1     2.14    2.14          2.14
#>  4 4       d1     4.32    4.32          4.32
#>  5 5       d1     3.57   NA             3.57
#>  6 6       d1     3.08    3.08          3.08
#>  7 7       d1     3.95    3.95          3.95
#>  8 8       d1     1.54    1.54          1.54
#>  9 9       d1     3.63    3.63          3.63
#> 10 10      d1     3.82   NA             3.82
#> # ... with 70 more rows

^{由 reprex package (v0.3.0)}

于 2020-06-07 创建

在 R 中使用 lag() 函数时用原始值替换 NA

replace NA with original value when using lag() function in R

r

lag

na

dplyr