r - 如何制作根据条件跨列多次连接的 R 循环

Question

我有一个类似于以下的数据框:

    df <- data.frame(name = c("john", "sara", "bill"),
                 join_0_year = c(1990,1991,1992),
                 join_0_month = c(1,2,3),
                 join_0_day = c(20,21,22),
                 join_0_team = c("tigers", "lions", "bears"),
                 join_1_year = c(2000, 2001, 2002),
                 join_1_month = c(8,9,10),
                 join_1_day = c(14,15,16),
                 join_1_team = c("pirates", "colts", "panthers"))
df

  name join_0_year join_0_month join_0_day join_0_team join_1_year join_1_month join_1_day join_1_team
1 john        1990            1         20      tigers        2000            8         14     pirates
2 sara        1991            2         21       lions        2001            9         15       colts
3 bill        1992            3         22       bears        2002           10         16    panthers

我的目标是将日期列连接成一个列。我的 df 有数百列(“join_0”到“join_400”)，因此按名称手动调用列是不可能的。我想我需要制作某种 for-loop，我为此苦苦挣扎。我想要的结果是:

  name join_0_date join_0_team join_1_date join_1_team
1 john   1990-1-20      tigers   2000-8-14     pirates
2 sara   1991-2-21       lions   2001-9-15       colts
3 bill   1992-3-22       bears  2002-10-16    panthers

非常感谢任何帮助。

Answer 1

这行得通，但比其他答案差得多。刚发现这个任务很有趣。

library(tidyverse) 

df %>% {lapply(seq(sum(str_detect(string = names(df), pattern = "day"))), 
                   function(y) y %>% {mutate(df, y=apply(df[(2+((.-1)*4)):(.*4)], 1, paste0, collapse = '-'))})} %>% 
      lapply(function(x) cbind(x$y)) %>% 
      cbind(df, .) %>% 
      dplyr::select(-which(str_detect(string = names(.), pattern = "year|month|day"))) %>% 
      setNames(ifelse(str_detect(names(.), pattern = "join_|name"), names(.), 
                      paste0("join_", seq(str_detect(names(.), pattern = "join_"))-length(.)+1, "_date")))

  name join_0_team join_1_team join_0_date join_1_date
1 john      tigers     pirates   1990-1-20   2000-8-14
2 sara       lions       colts   1991-2-21   2001-9-15
3 bill       bears    panthers   1992-3-22  2002-10-16