按照逻辑顺序删除数据框中的行答案

【问题标题】：Removing rows in a data frame following a logical sequence按照逻辑顺序删除数据框中的行
【发布时间】：2020-10-21 07:24:53
【问题描述】：

我希望有一个“扩张”后跟“收缩”的逻辑顺序，而不是紧密重复。

考虑这个数据框：

df <- structure(list(Date = structure(c(2922, 3652, 4018, 17897, 7305, 
8766, 13514, 14610, 14975), class = "Date"), LC = c(2148.55025151515, 
997.863792207792, 1027.36556255411, 5931.96483571429, -5317.829504329, 
-1293.82342294372, -1733.40640844156, -1558.7330974026, -2262.76932705628
), State = c("Expansion", "Contraction", "Expansion", "Expansion", 
"Contraction", "Contraction", "Contraction", "Expansion", "Contraction"
)), row.names = c(NA, -9L), class = "data.frame")

想要的输出是：

newdf <- structure(list(Date = structure(c(2922, 3652, 4018, 7305, 14610
), class = "Date"), LC = c(2148.55025151515, 997.863792207792, 
1027.36556255411, -5317.829504329, -1558.7330974026), State = c("Expansion", 
"Contraction", "Expansion", "Contraction", "Expansion")), row.names = c(1L, 
2L, 3L, 5L, 8L), class = "data.frame")

【问题讨论】：

标签： r dataframe dplyr conditional-statements

【解决方案1】：

您可以使用lead 和lag 来检查上一个和下一个值：

library(dplyr)

df %>%
  filter(State == 'Expansion' & lead(State) == 'Contraction' | 
         State == 'Contraction' & lag(State) == 'Expansion')

#        Date         LC       State
#1 1978-01-01  2148.5503   Expansion
#2 1980-01-01   997.8638 Contraction
#3 2019-01-01  5931.9648   Expansion
#4 1990-01-01 -5317.8295 Contraction
#5 2010-01-01 -1558.7331   Expansion
#6 2011-01-01 -2262.7693 Contraction

【讨论】：

【解决方案2】：

也许，我们可以filter 基于 'State' 的 run-length-id 上的重复项

library(dplyr)
library(data.table)
df %>%
      filter(!duplicated(rleid(State)))

【讨论】：