根据列的第一个值对所有行进行子集答案

【问题标题】：Subset All Rows Based on First Value of Column根据列的第一个值对所有行进行子集
【发布时间】：2020-04-27 10:18:59
【问题描述】：

我遇到了一个子集问题，我被困住了。这是数据的sn-p：

UniqueID MonthYear FirstObs
ABC123   OCT-18    1
ABC123   NOV-18    0
ABC123   JAN-19    0
ABC123   FEB-19    0
DEF446   MAY-19    1
DEF456   JUN-19    0
DEF456   JUL-19    0
GHI789   OCT-18    1
GHI789   NOV-18    0

数据集非常大，有一系列类似于上述示例的行。我希望能够编写一个子集公式，该公式提取具有相同 UniqueID 块的每一行，以 FirstObs=1 开头，并根据它们起源的月份将它们放在一起。我会有类似以下的内容：

Subset1 (all uniqueIDs that originated in October)
UniqueID MonthYear FirstObs
ABC123   OCT-18    1
ABC123   NOV-18    0
ABC123   JAN-19    0
ABC123   FEB-19    0
GHI789   OCT-18    1
GHI789   NOV-18    0


Subset2 (all uniqueIDs that originated in May)
UniqueID MonthYear FirstObs    
DEF446   MAY-19    1
DEF456   JUN-19    0
DEF456   JUL-19    0

理想情况下，每个月以 FirstObs=1 开头的每个块都有一个子集。我知道我需要使用一些 ifelse 和子集函数系列，但我不确定如何以最佳方式使用它们。

【问题讨论】：

标签： r if-statement subset

【解决方案1】：

这可能有帮助

library(dplyr)
df2 <- df1 %>% 
     group_by(UniqueID) %>% 
     filter(first(FirstObs) == 1 & n() > 1)
df3 <- anti_join(df1, df2)

也可以

library(stringr)
df2 <- df1 %>%
          group_by(UniqueID) %>% 
          filter(first(FirstObs) == 1, 
                str_remove(first(MonthYear), "-\\d+") == "OCT")
df3 <- anti_join(df1, df2)

或将split 转换为data.frames 的list

df1 %>%
   group_by(UniqueID) %>%
   mutate(grp = first(FirstObs) == 1 & n() > 1) %>%
   ungroup %>%
   group_split(grp, keep = FALSE)
#[[1]]
# A tibble: 3 x 3
#  UniqueID MonthYear FirstObs
#  <chr>    <chr>        <int>
#1 DEF446   MAY-19           1
#2 DEF456   JUN-19           0
#3 DEF456   JUL-19           0

#[[2]]
# A tibble: 6 x 3
#  UniqueID MonthYear FirstObs
#  <chr>    <chr>        <int>
#1 ABC123   OCT-18           1
#2 ABC123   NOV-18           0
#3 ABC123   JAN-19           0
#4 ABC123   FEB-19           0
#5 GHI789   OCT-18           1
#6 GHI789   NOV-18           0

数据

df1 <- structure(list(UniqueID = c("ABC123", "ABC123", "ABC123", "ABC123", 
"DEF446", "DEF456", "DEF456", "GHI789", "GHI789"), MonthYear = c("OCT-18", 
"NOV-18", "JAN-19", "FEB-19", "MAY-19", "JUN-19", "JUL-19", "OCT-18", 
"NOV-18"), FirstObs = c(1L, 0L, 0L, 0L, 1L, 0L, 0L, 1L, 0L)), 
class = "data.frame", row.names = c(NA, 
-9L))

【讨论】：