[英]Create new variable using mutate on logical conditions across many variables - mutate?
[英]Is there a way to mutate and create many new variables with one line of code?
我有類似的東西:
df<-data.frame(group=c(1, 1, 1, 1,1,2, 2, 2, 2),
date=c("2000-01-01 11:00:00", "2000-01-03 11:00:00", "2000-01-04 11:20:00", "2000-01-04 14:20:00", "2000-01-05 11:40:00", "2000-01-09 12:20:00", "2000-01-09 13:20:00", "2000-01-10 12:20:00", "2000-01-12 16:20:00"))
group date
1 1 2000-01-01 11:00:00
2 1 2000-01-03 11:00:00
3 1 2000-01-04 11:20:00
4 1 2000-01-04 14:20:00
5 1 2000-01-05 11:40:00
6 2 2000-01-09 12:20:00
7 2 2000-01-09 13:20:00
8 2 2000-01-10 12:20:00
9 2 2000-01-12 16:20:00
我想制作許多列指示日期后 24 小時、48 小時等(例如):
df%>%mutate(first=date+86400, second=date+172800, third=date+259200)
等等等等,我在幾秒鍾內添加一天,但這非常耗時(如果我想要數百列)。 我假設有一種方法可以迭代地執行此操作。
謝謝,
如果我們可以使用english
包,則可以生成列名,而使用lapply
生成值
library(english)
df$date <- as.POSIXct(df$date)
df[as.character(ordinal(1:3))] <- lapply(1:3, function(x) df$date + 86400 * x)
它可以在循環中使用as.POSIXct
轉換在一行代碼中完成,但我們將不必要地多次進行轉換(並非所有單行代碼都是有效的)
或者帶着purrr
library(purrr)
library(dplyr)
map_dfc(1:3, ~ tibble(!! as.character(ordinal(.x)) := df$date + 86400 * .x)) %>%
bind_cols(df, .)
# group date first second third
#1 1 2000-01-01 11:00:00 2000-01-02 11:00:00 2000-01-03 11:00:00 2000-01-04 11:00:00
#2 1 2000-01-03 11:00:00 2000-01-04 11:00:00 2000-01-05 11:00:00 2000-01-06 11:00:00
#3 1 2000-01-04 11:20:00 2000-01-05 11:20:00 2000-01-06 11:20:00 2000-01-07 11:20:00
#4 1 2000-01-04 14:20:00 2000-01-05 14:20:00 2000-01-06 14:20:00 2000-01-07 14:20:00
#5 1 2000-01-05 11:40:00 2000-01-06 11:40:00 2000-01-07 11:40:00 2000-01-08 11:40:00
#6 2 2000-01-09 12:20:00 2000-01-10 12:20:00 2000-01-11 12:20:00 2000-01-12 12:20:00
#7 2 2000-01-09 13:20:00 2000-01-10 13:20:00 2000-01-11 13:20:00 2000-01-12 13:20:00
#8 2 2000-01-10 12:20:00 2000-01-11 12:20:00 2000-01-12 12:20:00 2000-01-13 12:20:00
#9 2 2000-01-12 16:20:00 2000-01-13 16:20:00 2000-01-14 16:20:00 2000-01-15 16:20:00
涉及dplyr
和purrr
另一種可能性可能是:
map(86400*1:3, ~ df %>%
transmute(!!paste(.x/3600, "hours", sep = "_") := as.POSIXct(date) + .x)) %>%
bind_cols(df, .)
group date 24_hours 48_hours 72_hours
1 1 2000-01-01 11:00:00 2000-01-02 11:00:00 2000-01-03 11:00:00 2000-01-04 11:00:00
2 1 2000-01-03 11:00:00 2000-01-04 11:00:00 2000-01-05 11:00:00 2000-01-06 11:00:00
3 1 2000-01-04 11:20:00 2000-01-05 11:20:00 2000-01-06 11:20:00 2000-01-07 11:20:00
4 1 2000-01-04 14:20:00 2000-01-05 14:20:00 2000-01-06 14:20:00 2000-01-07 14:20:00
5 1 2000-01-05 11:40:00 2000-01-06 11:40:00 2000-01-07 11:40:00 2000-01-08 11:40:00
6 2 2000-01-09 12:20:00 2000-01-10 12:20:00 2000-01-11 12:20:00 2000-01-12 12:20:00
7 2 2000-01-09 13:20:00 2000-01-10 13:20:00 2000-01-11 13:20:00 2000-01-12 13:20:00
8 2 2000-01-10 12:20:00 2000-01-11 12:20:00 2000-01-12 12:20:00 2000-01-13 12:20:00
9 2 2000-01-12 16:20:00 2000-01-13 16:20:00 2000-01-14 16:20:00 2000-01-15 16:20:00
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.