需要帮助重塑 R 数据集答案

【问题标题】：Need help reshaping an R dataset需要帮助重塑 R 数据集
【发布时间】：2021-01-06 18:46:28
【问题描述】：

我的数据集目前看起来像这样：

id  date00  var1_00 var2_00 date01  var1_01 var2_01
1   1/1/2019    1   2       1/1/2020    3   4
2   2/2/2019    1   2       2/2/2020    3   4
3   3/3/2019    1   2       3/3/2020    3   4

表格代码：

structure(list(id = c(1, 2, 3), date00 = structure(c(1546300800, 
1549065600, 1551571200), class = c("POSIXct", "POSIXt"), tzone = "UTC"), 
    var1_00 = c(1, 1, 1), var2_00 = c(2, 2, 2), date01 = structure(c(1577836800, 
    1580601600, 1583193600), class = c("POSIXct", "POSIXt"), tzone = "UTC"), 
    var1_01 = c(3, 3, 3), var2_01 = c(4, 4, 4)), row.names = c(NA, 
-3L), class = c("tbl_df", "tbl", "data.frame"))

我怎样才能重塑它，使它看起来像这样：

id  date    var1_00 var2_00 var1_01 var2_01
1   1/1/2019    1   2       NA      NA
2   2/2/2019    1   2       NA      NA
3   3/3/2019    1   2       NA      NA
1   1/1/2020    NA  NA      3       4
2   1/1/2020    NA  NA      3       4
3   1/1/2020    NA  NA      3       4

谢谢！

【问题讨论】：

您能否提供用于创建示例表的代码？使用原始的data.frame() 调用，或使用dput() 生成代码
给你：> dput(Example) structure(list(id = c(1, 2, 3), date00 = structure(c(1546300800, 1549065600, 1551571200), class= c(" POSIXct", "POSIXt"), tzone = "UTC"), var1_00 = c(1, 1, 1), var2_00 = c(2, 2, 2), date01 = structure(c(1577836800, 1580601600, 1583193600), class= c("POSIXct", "POSIXt"), tzone = "UTC"), var1_01 = c(3, 3, 3), var2_01 = c(4, 4, 4)), row.names = c(NA , -3L), class= c("tbl_df", "tbl", "data.frame"))

标签： r dataframe reshape melt

【解决方案1】：

这是一个使用rbindlist 的data.table 选项

setDT(df)
dt1 <- setnames(df[,.SD,.SDcols = grep("^id|00$",names(df))],"date00","date")
dt2 <- setnames(df[,.SD,.SDcols = grep("^id|01$",names(df))],"date01","date")
out <- rbindlist(list(dt1,dt2),fill = TRUE)

或

dt <- as.data.table(df)
out <- rbindlist(
  lapply(
    split.default(dt[,-1],gsub(".*(\\d+$)","\\1",names(dt)[-1])),
    function(x) cbind(dt[,1],setnames(x,1,"date"))),
  fill = TRUE
)

这样

> out
   id       date var1_00 var2_00 var1_01 var2_01
1:  1 2019-01-01       1       2      NA      NA
2:  2 2019-02-02       1       2      NA      NA
3:  3 2019-03-03       1       2      NA      NA
4:  1 2020-01-01      NA      NA       3       4
5:  2 2020-02-02      NA      NA       3       4
6:  3 2020-03-03      NA      NA       3       4

【讨论】：

【解决方案2】：

我尝试了一些并得出了这个解决方案。请告诉我。

library(dplyr)

df1 <- df %>% 
  mutate(date=date00,
         var1_01=NA,
         var2_01=NA) %>% 
  select(id, date, var1_00, var2_00, var1_01, var2_01) 

df2 <- df %>% 
  mutate(date=date01,
         var1_00=NA,
         var2_00=NA) %>% 
  select(id, date, var1_00, var2_00, var1_01, var2_01)

df_new <- rbind(df1, df2)

【讨论】：