R melting/gathering 包含 2 个标题行到属性中的文件

Question

我需要读入一个 2 级 headers 的数据文件，数据如下所示：

|          | Jone Doe |      |      |      |      |      |      | Jane Doe |      |      |      |      |      |      | 
|----------|----------|------|------|------|------|------|------|----------|------|------|------|------|------|------| 
| Date     | Col1     | Col2 | Col3 | Col4 | Col5 | Col6 | Col7 | Col1     | Col2 | Col3 | Col4 | Col5 | Col6 | Col7 | 
| 1-Jul-13 | 49       | 42   | 20   | 18   | 23   | 16   | 29   | 48       | 33   | 24   | 10   | 43   | 13   | 43   | 
| 2-Jul-13 | 17       | 16   | 43   | 33   | 37   | 37   | 10   | 7        | 45   | 19   | 4    | 41   | 41   | 20   | 
| 3-Jul-13 | 35       | 39   | 42   | 35   | 5    | 12   | 22   | 3        | 28   | 23   | 10   | 12   | 5    | 8    |

我需要它看起来像这样：

| Date     | Name     | Col1 | Col2 | Col3 | Col4 | Col5 | Col6 | Col7 | 
|----------|----------|------|------|------|------|------|------|------| 
| 1-Jul-13 | Jone Doe | 49   | 42   | 20   | 18   | 23   | 16   | 29   | 
| 2-Jul-13 | Jone Doe | 17   | 16   | 43   | 33   | 37   | 37   | 10   | 
| 3-Jul-13 | Jone Doe | 35   | 39   | 42   | 35   | 5    | 12   | 22   | 
| 1-Jul-13 | Jane Doe | 48   | 33   | 24   | 10   | 43   | 13   | 43   | 
| 2-Jul-13 | Jane Doe | 7    | 45   | 19   | 4    | 41   | 41   | 20   | 
| 3-Jul-13 | Jane Doe | 3    | 28   | 23   | 10   | 12   | 5    | 8    |

知道如何在没有 hard-coding 的情况下执行此操作吗？我一直在尝试使用 melt() 和 gather() 没有任何运气

编辑：

示例数据：https://drive.google.com/open?id=1T4KkAk5D55_nXsHlr1Aozed6d49qFM_8

lst1 的输出：

nm1 的输出：

 [1] "John Doe"  "John Doe"  "John Doe"  "John Doe"  "John Doe"  "John Doe"  "John Doe"  "Jane Doe" 
 [9] "Jane Doe"  "Jane Doe"  "Jane Doe"  "Jane Doe"  "Jane Doe"  "Jane Doe"  "Jose Doe"  "Jose Doe" 
[17] "Jose Doe"  "Jose Doe"  "Jose Doe"  "Jose Doe"  "Jose Doe"  "Jacob Doe" "Jacob Doe" "Jacob Doe"
[25] "Jacob Doe" "Jacob Doe" "Jacob Doe" "Jacob Doe"

Answer 1

一个选项是使用 skip 参数读取数据集以跳过第一行，然后，我们可以根据重复的列名 split 将数据 list然后根据第一行和 rbind list 元素为每个 list 元素创建 'Name' 列以创建单个 data.frame

dat1 <- read.csv("file.csv", header = TRUE, skip = 1, 
          stringsAsFactors = FALSE, na.strings = "N/A")
nm1 <- c("John Doe", "Jane Doe")[cumsum(grepl("Col1", names(dat1)[-1]))]
 nm2 <- unique( sub("\.\d+$", "", names(dat1)[-1]))
lst1 <- split.default(dat1[-1], nm1)
dat2 <- cbind(dat1['Date'], do.call(rbind, Map(cbind, Name = nm1,  lapply(lst1, setNames, nm2))))
row.names(dat2) <- NULL
head(dat2, 5)
#      Date     Name Col1 Col2 Col3 Col4 Col5 Col6 Col7
#1 1-Jul-13 John Doe   52    6   NA   NA    7   20   25
#2 2-Jul-13 John Doe   43    7   NA   NA   NA   25   17
#3 3-Jul-13 John Doe   55    5   NA   NA    4   23   28
#4 4-Jul-13 John Doe   42    6   NA   NA    7   21   14
#5 5-Jul-13 John Doe   64    3   NA   NA    5   36   22

dim(dat2)
#[1] 140   9

注意，如果列的块数很大，一个选项是用readLines

读取第一行

v1 <- readLines("file.csv", n = 1)
v2 <- scan(text = gsub(",{2,}", ",", trimws(v1)), sep=",", what = "", quiet = TRUE)
v3 <- v2[nzchar(v2)]

并将其提供给 cumsum 步骤

nm1 <- v3[cumsum(grepl("Col1", names(dat1)[-1]))]

R melting/gathering 包含 2 个标题行到属性中的文件

R melting/gathering file with 2 heading rows into attributes

r

melt

tidyverse