我有一个数据框,我想在其中检查一些条件并需要根据条件结果添加一个新列。
这是我的输入数据
InputData = data.frame(A = c("", "", "Apple"), B = c("", "", "Orange"), C = c("", "", ""), D = c(0, 1, 1))
这是我想要的输出
OutputData = InputData %>%
mutate(R = case_when(A=='' & B=='' & C=='' & D==0 ~ "Yes", TRUE ~ "No"))
我尝试使用 Case 函数进行变异。它工作正常但是当我有更多行数时需要更长的时间。
请帮助我以更快的方式完成它。
我很惊讶你的代码对于这么小的数据(只有 100k 行)很慢。我会这样做:
InputData$R <- "No"
InputData[InputData$A == '' & InputData$B == '' &
InputData$C == '' & InputData$D == 0, "R"] <- "Yes"
但是,我强烈建议使用逻辑值而不是“是”/“否”:
InputData$S <- InputData$A == '' & InputData$B == '' &
InputData$C == '' & InputData$D == 0
# A B C D R S
#1 0 Yes TRUE
#2 1 NO FALSE
#3 Apple Orange 1 NO FALSE
如果仍然太慢,包 data.table 可以提供帮助。但除非数据实际上变大,否则没有必要这样做。
下面是一个简单的基准。看来如果只是想二分列,
if_else
比case_when
中的dplyr
更可取。如果您关心速度,请将工作流程更改为base
像@Roland的回答。
InputData = data.frame(A = sample(c('x', ''), 1e5, TRUE),
B = sample(c('x', ''), 1e5, TRUE),
C = sample(c('x', ''), 1e5, TRUE),
D = sample(0:1, 1e5, TRUE))
library(dplyr)
bench::mark(
"base::ifelse" = InputData %>% mutate(R = ifelse(A == '' & B == '' & C == '' & D == 0, "Yes", "No")),
"dplyr::case_when" = InputData %>% mutate(R = case_when(A == '' & B == '' & C == '' & D == 0 ~ "Yes", TRUE ~ "No")),
"dplyr::if_else" = InputData %>% mutate(R = if_else(A == '' & B == '' & C == '' & D == 0, "Yes", "No")),
"base::repalce" = InputData %>% mutate(R = "No", R = replace(R, A == '' & B == '' & C == '' & D == 0, "Yes")),
"base::`[<-`.Roland" = local({
InputData$R <- "No"
InputData$R[InputData$A == '' & InputData$B == '' & InputData$C == '' & InputData$D == 0] <- "Yes"
InputData
}),
iterations = 100
)
# # A tibble: 5 × 9
# expression min median `itr/sec` mem_alloc `gc/sec` n_itr n_gc total_time
# <bch:expr> <bch:tm> <bch:tm> <dbl> <bch:byt> <dbl> <int> <dbl> <bch:tm>
# 1 base::ifelse 24.87ms 25.82ms 38.0 7.63MB 17.1 69 31 1.82s
# 2 dplyr::case_when 15.65ms 16.91ms 57.0 8.4MB 24.4 70 30 1.23s
# 3 dplyr::if_else 6.77ms 7.17ms 133. 6.87MB 39.6 77 23 580.57ms
# 4 base::repalce 5.6ms 5.9ms 166. 5.75MB 36.4 82 18 495.1ms
# 5 base::`[<-`.Roland 3.47ms 3.52ms 269. 3.84MB 33.2 89 11 331.35ms