根据组中第一个值的条件替换 df 中的后续值

Question

我在有序的 R 数据框中有这种类型的数据。

set.seed(25)

date <- sort(as.Date(sample( as.numeric(as.Date("2019-01-01")): as.numeric(as.Date("2021-03-31")), 10, 
                             replace = T), 
                     origin = '1970-01-01'))

type <- c("Football", "Football", "Rugby", "Football", "Hockey", "Tennis", "Hockey", "Basketball", "Basketball", "Rugby")

id <- c("1","1","1","1","2","2","3","4","4","5")

df <- data.frame(date,id, type)



      date id       type
  2019-04-09  1   Football
  2019-04-13  1   Football
  2019-04-20  1      Rugby
  2019-04-21  1   Football
  2019-05-31  2     Hockey
  2020-02-09  2     Tennis
  2020-03-08  3     Hockey
  2020-03-24  4 Basketball
  2020-08-18  4   Football
  2020-11-01  5      Rugby

我试图得到的结果是这样的：

    date id       type     type_2
  2019-04-09  1   Football   Football
  2019-04-13  1   Football   Football
  2019-04-20  1      Rugby      Multi
  2019-04-21  1   Football      Multi
  2019-05-31  2     Hockey     Hockey
  2020-02-09  2     Tennis      Multi
  2020-03-08  3     Hockey     Hockey
  2020-03-24  4 Basketball Basketball
  2020-08-18  4 Basketball Basketball
  2020-11-01  5      Rugby      Rugby

基本上，如果他练习的下一项运动与前一项运动相同，则id练习的第一项运动保持不变，type_2保持不变，但一旦他稍后改变运动，他稍后将他的其余值更改为 multi。

我尝试在 dplyr 中使用 lag()、lead() 和 if_else() 执行此操作，但结果永远不会如我所愿。

Answer 1

您可以使用 data.table 中的 rleid 为每个 id 中的 type 变量生成运行长度 ID。第一次更改后的所有内容变为 "Multi".

library(data.table)

setDT(df)[, type2 := replace(type, rleid(type) > 1, 'Multi'), id]
df

#          date id       type      type2
# 1: 2019-02-18  1   Football   Football
# 2: 2019-02-28  1   Football   Football
# 3: 2019-03-13  1      Rugby      Multi
# 4: 2019-09-29  1   Football      Multi
# 5: 2019-10-09  2     Hockey     Hockey
# 6: 2020-03-19  2     Tennis      Multi
# 7: 2020-04-21  3     Hockey     Hockey
# 8: 2020-06-19  4 Basketball Basketball
# 9: 2020-09-08  4 Basketball Basketball
#10: 2020-10-08  5      Rugby      Rugby

如果你喜欢写成dplyr-

library(dplyr)

df %>%
  group_by(id) %>%
  mutate(type2 = replace(type, rleid(type) > 1, 'Multi')) %>%
  ungroup

根据组中第一个值的条件替换 df 中的后续值

Replace subsequent values in a df based on condition of first value in group

r

lag

lead