将多个 header table 转换为长格式

Question

我正在阅读包含多行 header 的 Excel table，它通过 read.csv 创建一个像这样的 object R.

R1 <- c("X", "X.1", "X.2", "X.3", "EU", "EU.1", "EU.2", "US", "US.1", "US.2")
R2 <- c("Min Age", "Max Age", "Min Duration", "Max Duration", "1", "2", "3", "1", "2", "3")
R3 <- c("18", "21", "1", "3", "0.12", "0.32", "0.67", "0.80", "0.90", "1.01")
R4 <- c("22", "25", "1", "3", "0.20", "0.40", "0.70", "0.85", "0.98", "1.05")
R5 <- c("26", "30", "1", "3", "0.25", "0.50", "0.80", "0.90", "1.05", "1.21")
R6 <- c("18", "21", "4", "5", "0.32", "0.60", "0.95", "0.99", "1.30", "1.40")
R7 <- c("22", "25", "4", "5", "0.40", "0.70", "1.07", "1.20", "1.40", "1.50")
R8 <- c("26", "30", "4", "5", "0.55", "0.80", "1.09", "1.34", "1.67", "1.99")
table1 <- as.data.frame(rbind(R1, R2, R3, R4, R5, R6, R7, R8))

我现在要如何 'flatten' 这样我才能得到 R table 和 "Min age"、"Max Age" , "Min Duration", "Max Duration", "Area", "Level", "Price" 列。 "Area" 列显示 "EU" 或 "US"，"Level" 列显示 1、2 或 3，然后 "Price" 列显示相应的价格在 Excel table?

中找到

如果没有多个 header 行，我会使用 tidyr 的收集函数，但似乎无法处理这些数据，有什么想法吗？

输出总共应该有 36 行 + headers

Answer 1

如果您按照 akrun 的建议跳过第一行，您可能会得到如下所示的数据：（自动添加 "X"s 和“.1”/“.2”通过 R)

library(tidyverse)

df <- tribble(
    ~Min.Age, ~Max.Age, ~Min.Duration, ~Max.Duration,  ~X1.1,  ~X2.1,  ~X3.1, ~X1.2, ~X2.2, ~X3.2,
    "18",     "21",           "1",           "3", "0.12", "0.32", "0.67",  "0.80",  "0.90",  "1.01",
    "22",     "25",           "1",           "3", "0.20", "0.40", "0.70",  "0.85",  "0.98",  "1.05",
    "26",     "30",           "1",           "3", "0.25", "0.50", "0.80",  "0.90",  "1.05",  "1.21",
    "18",     "21",           "4",           "5", "0.32", "0.60", "0.95",  "0.99",  "1.30",  "1.40",
    "22",     "25",           "4",           "5", "0.40", "0.70", "1.07",  "1.20",  "1.40",  "1.50",
    "26",     "30",           "4",           "5", "0.55", "0.80", "1.09",  "1.34",  "1.67",  "1.99"
)

有了这些数据，您就可以使用 gather 将所有以 X 开头的 headers 收集到一列中，并将价格收集到另一列中。你可以separate把headers变成"Level"和"Area"。最后，重新编码 Area 并从关卡中移除 "X"。

df %>% 
    gather(headers, Price, starts_with("X")) %>% 
    separate(headers, c("Level", "Area")) %>% 
    mutate(Area = if_else(Area == "1", "EU", "US"),
           Level = parse_number(Level))
#> # A tibble: 36 x 7
#>    Min.Age Max.Age Min.Duration Max.Duration Level Area  Price
#>    <chr>   <chr>   <chr>        <chr>        <dbl> <chr> <chr>
#>  1 18      21      1            3                1 EU    0.12 
#>  2 22      25      1            3                1 EU    0.20 
#>  3 26      30      1            3                1 EU    0.25 
#>  4 18      21      4            5                1 EU    0.32 
#>  5 22      25      4            5                1 EU    0.40 
#>  6 26      30      4            5                1 EU    0.55 
#>  7 18      21      1            3                2 EU    0.32 
#>  8 22      25      1            3                2 EU    0.40 
#>  9 26      30      1            3                2 EU    0.50 
#> 10 18      21      4            5                2 EU    0.60 
#> # ... with 26 more rows

^{由 reprex package (v0.2.1)}

创建于 2018-10-12

P.S。您可以在此处找到大量电子表格修改工作流程：https://nacnudus.github.io/spreadsheet-munging-strategies/small-multiples-with-all-headers-present-for-each-multiple.html

将多个 header table 转换为长格式

Convert multiple header table to long format

excel

r

flatten