使用 dplyr 和 tidyr 计算分组 data.frame 的平均值

Question

我只是在学习 R 并试图找到方法来修改我的分组 data.frame 以获得变量 value (x+y/2) 和标准的平均值内聚观察的偏差 (sd) sqrt((x^2+y^2)/2)。其他（相等）变量（sequence、value1）不应更改。

我使用了 subset() 和 rowMeans()，但我想知道使用 dplyr 和 tidyr 是否有更好的方法（可能使用嵌套数据框？）

我的测试 data.frame 看起来像：

id      location    value  sd    sequence value1
"anon1" "nose"      5      0.2    "a"      1
"anon2" "body"      4      0.4    "a"      2
"anon3" "left_arm"  3      0.3    "a"      3
"anon3" "right_arm" 5      0.6    "a"      3
"anon4" "head"      4      0.3    "a"      4
"anon5" "left_leg"  2      0.2    "a"      5
"anon5" "right_leg" 1      0.1    "a"      5

dput 我测试的输出 data.frame:

myData <- structure(list(ï..id = structure(c(1L, 2L, 3L, 3L, 4L, 5L, 5L
), .Label = c("anon1", "anon2", "anon3", "anon4", "anon5"), class = "factor"), 
    location = structure(c(5L, 1L, 3L, 6L, 2L, 4L, 7L), .Label = c("body", 
    "head", "left_arm", "left_leg", "nose", "right_arm", "right_leg"
    ), class = "factor"), value = c(5L, 4L, 3L, 5L, 4L, 2L, 1L
    ), sd = c(0.2, 0.4, 0.3, 0.6, 0.3, 0.2, 0.1), sequence = structure(c(1L, 
    1L, 1L, 1L, 1L, 1L, 1L), .Label = "a", class = "factor"), 
    value1 = c(1L, 2L, 3L, 3L, 4L, 5L, 5L)), .Names = c("ï..id", 
"location", "value", "sd", "sequence", "value1"), class = "data.frame", row.names = c(NA, 
-7L))

它应该是什么样子：

id      location value sd   sequence value1
"anon1" "nose"   5     0.2  "a"      1 
"anon2" "body"   4     0.4  "a"      2
"anon3" "arm"    4     0.47 "a"      3
"anon4" "head"   4     0.3  "a"      4
"anon5" "leg"    1.5   0.15 "a"      5

Answer 1

dplyr 的 group_by 和 summarise 会有所帮助，gsub 对字符串变量提供了一些支持：

library(dplyr)

myData %>% 
  group_by(id) %>% 
  summarise(
    location = gsub(".*_", "", location[1]),
    value = mean(value),
    sd = mean(sd),
    sequence = sequence[1],
    value1 = value1[1]
  )
#> # A tibble: 5 × 6
#>       id location value    sd sequence value1
#>   <fctr>    <chr> <dbl> <dbl>   <fctr>  <int>
#> 1  anon1     nose   5.0  0.20        a      1
#> 2  anon2     body   4.0  0.40        a      2
#> 3  anon3      arm   4.0  0.45        a      3
#> 4  anon4     head   4.0  0.30        a      4
#> 5  anon5      leg   1.5  0.15        a      5

或者如果 id、sequence 和 value1 在所有情况下都匹配：

myData %>% 
  group_by(id, sequence, value1) %>% 
  summarise(
    location = gsub(".*_", "", location[1]),
    value = mean(value),
    sd = mean(sd))
#> Source: local data frame [5 x 6]
#> Groups: id, sequence [?]
#> 
#>       id sequence value1 location value    sd
#>   <fctr>   <fctr>  <int>    <chr> <dbl> <dbl>
#> 1  anon1        a      1     nose   5.0  0.20
#> 2  anon2        a      2     body   4.0  0.40
#> 3  anon3        a      3      arm   4.0  0.45
#> 4  anon4        a      4     head   4.0  0.30
#> 5  anon5        a      5      leg   1.5  0.15

使用 dplyr 和 tidyr 计算分组 data.frame 的平均值

Compute average of grouped data.frame using dplyr and tidyr

average

r

nested-datalist

dplyr

tidyr

我的测试 data.frame 看起来像：

dput 我测试的输出 data.frame:

它应该是什么样子：