在R中格式化数据帧

我有一个相当复杂的任务，我需要执行，所以请容忍我。我猜这是可能的，但如果不可能，请告诉我

假设我有以下数据

set.seed(123)
date1 <- c(seq(as.Date("2011-11-1"),as.Date("2012-1-1"),by = "months"),seq(as.Date("2011-12-1"),as.Date("2012-3-1"),by = "months"))
date2 <- c(seq(as.Date("2011-12-1"),as.Date("2012-1-1"),by = "months"),seq(as.Date("2011-11-1"),as.Date("2012-1-1"),by = "months"))
variables <- c(rep("Number of Coins",3),rep("Number of Shoes",4),rep("Number of Coins",2),rep("Number of Shoes",3))
date <- c(date1,date2)
names <- c(rep("Jim",7),rep("Arnold",5))
value <- rnorm(12)
df <- data.frame(names, date, variables, value)

    names       date       variables       value
1     Jim 2011-11-01 Number of Coins -0.56047565
2     Jim 2011-12-01 Number of Coins -0.23017749
3     Jim 2012-01-01 Number of Coins  1.55870831
4     Jim 2011-12-01 Number of Shoes  0.07050839
5     Jim 2012-01-01 Number of Shoes  0.12928774
6     Jim 2012-02-01 Number of Shoes  1.71506499
7     Jim 2012-03-01 Number of Shoes  0.46091621
8  Arnold 2011-12-01 Number of Coins -1.26506123
9  Arnold 2012-01-01 Number of Coins -0.68685285
10 Arnold 2011-11-01 Number of Shoes -0.44566197
11 Arnold 2011-12-01 Number of Shoes  1.22408180
12 Arnold 2012-01-01 Number of Shoes  0.35981383

---

因此，日期范围将是每个变量的最小日期到每个变量的最大日期。这将产生对NAs的需求。我想在每个

名称中执行此操作。希望这有意义
 正如@Ajinkya Kale建议的那样，您可以使用
reformae2
包处理此任务

dcast(df, names + date ~ variables, value.var = "value")

arrange(dcast(df, names + date ~ variables, value.var = "value"), names, date)

#   names       date Number of Coins Number of Shoes
#1 Arnold 2011-11-01              NA     -0.44566197
#2 Arnold 2011-12-01      -1.2650612      1.22408180
#3 Arnold 2012-01-01      -0.6868529      0.35981383
#4    Jim 2011-11-01      -0.5604756              NA
#5    Jim 2011-12-01      -0.2301775      0.07050839
#6    Jim 2012-01-01       1.5587083      0.12928774
#7    Jim 2012-02-01              NA      1.71506499
#8    Jim 2012-03-01              NA      0.46091621

如果要确保日期顺序是按时间顺序排列的，可以在
dplyr
包中使用
arrange（）

dcast(df, names + date ~ variables, value.var = "value")

arrange(dcast(df, names + date ~ variables, value.var = "value"), names, date)

#   names       date Number of Coins Number of Shoes
#1 Arnold 2011-11-01              NA     -0.44566197
#2 Arnold 2011-12-01      -1.2650612      1.22408180
#3 Arnold 2012-01-01      -0.6868529      0.35981383
#4    Jim 2011-11-01      -0.5604756              NA
#5    Jim 2011-12-01      -0.2301775      0.07050839
#6    Jim 2012-01-01       1.5587083      0.12928774
#7    Jim 2012-02-01              NA      1.71506499
#8    Jim 2012-03-01              NA      0.46091621

另一种选择是使用
tidyr

library(tidyr)
spread(df, variables, value)
#   names       date Number of Coins Number of Shoes
#1 Arnold 2011-11-01              NA     -0.44566197
#2 Arnold 2011-12-01      -1.2650612      1.22408180
#3 Arnold 2012-01-01      -0.6868529      0.35981383
#4    Jim 2011-11-01      -0.5604756              NA
#5    Jim 2011-12-01      -0.2301775      0.07050839
#6    Jim 2012-01-01       1.5587083      0.12928774
#7    Jim 2012-02-01              NA      1.71506499
#8    Jim 2012-03-01              NA      0.46091621

您可以使用重塑2的dcast melt