如何根据 R 中的一列从另一列和行制作新的列复制值？

Question

这是一些示例数据：

data <- tibble(line_number = seq(1:5),
               test = c("testA", "testB", "testC", "testD", "testE"),
               start_date = as_date(c("2021-01-01", "2021-02-01", "2021-02-15", "2021-03-20", "2021-04-12")),
               finish_date = as_date(c("2021-01-01", "2021-03-01", "2021-02-18", "2021-05-20", "2021-04-12")),
               coded_date = c(NA, "1S", "2F", "2S", "4F"))

#       line_number test  start_date finish_date coded_date
#             <int> <chr> <date>     <date>      <chr>     
# 1               1 testA 2021-01-01 2021-01-01  NA        
# 2               2 testB 2021-02-01 2021-03-01  1S        
# 3               3 testC 2021-02-15 2021-02-18  2F        
# 4               4 testD 2021-03-20 2021-05-20  2S        
# 5               5 testE 2021-04-12 2021-04-12  4F

我想创建两个名为“new_start_date”和“new_finish_date”的新列，其内容由“coded_date”列确定。

对于coded_date列：数字对应行号，“S”表示开始日期，“F”表示结束日期。

以第2行1S为例，我希望“new_start_date”列取第1行开始日期的值并复制，留下“new_finish_date" 与 NA.

以第3行2F为例，我希望“new_finish_date”列取第2行的完成日期值并复制，留下“new_start_date" 与 NA.

这是我想要的输出：

#       line_number test  start_date finish_date coded_date new_start_date new_finish_date
#             <int> <chr> <date>     <date>      <chr>      <date>         <date>
# 1               1 testA 2021-01-01 2021-01-01  NA         NA             NA
# 2               2 testB 2021-02-01 2021-03-01  1S         2021-01-01     NA
# 3               3 testC 2021-02-15 2021-02-18  2F         NA             2021-03-01
# 4               4 testD 2021-03-20 2021-05-20  2S         2021-02-01     NA
# 5               5 testE 2021-04-12 2021-04-12  4F         NA             2021-05-20

我仍然是使用 R 的初学者，因此非常感谢任何帮助或输入:)

Answer 1

我的第一直觉是使用两个 for 循环

# deal with start date first
# get all values in 'coded_date' that contain an 'S'
svals <- grep(pattern = "S", x = data$coded_date, value = TRUE)
# we'll go row by row
# for each case (row) in that contains an 'S'
for(sval in svals){
  # get the rowid from the value of 'coded_date'
  # this is the row where we'll get the new date
  rowid <- substring(text = sval, first = 1, last = 1)
  # assign a 'new_start_date' to the row where we found sval
  # the row containing this new value is defined by rowid
  # use '%in% rather than '==' on left side because NAs are present
  data[data$coded_date %in% sval, "new_start_date"] <- data[rowid,"start_date"]
}

## repeat for finish date
# S and F loops could be nested together!
fvals <- grep(pattern = "F", x = data$coded_date, value = TRUE)
for(fval in fvals){
  rowid <- substring(text = fval, first = 1, last = 1)
  data[data$coded_date %in% fval, "new_finish_date"] <- data[rowid,"finish_date"]

e：这是一个似乎有效的矢量化版本。如果有人热衷于此，我敢打赌这里还有改进的余地，我将不胜感激任何反馈！

data$new_start_date2 <- ifelse(
  test = grepl(pattern = "S", x = data$coded_date),
  yes = data[sub(pattern = "S", replacement = "", data$coded_date),"start_date"],
  no = NA)
data$new_finish_date2 <- ifelse(
  test = grepl(pattern = "F", x = data$coded_date),
  yes = data[sub(pattern = "F", replacement = "", data$coded_date),"finish_date"],
  no = NA)

如何根据 R 中的一列从另一列和行制作新的列复制值？

How to make new column copy values from another column AND row, based on one column in R?

r

date

calculated-columns

dataframe

tibble