删除行 w。重复信息，同时保留第一个非重复条目（并将重复条目中的数据附加到该行）

Question

我有一个包含 50 万个条目（行）的数据集。每个条目都是针对特定学生的，并包含有关该学生所在学校的信息去了那个特定的学期。

因为学生在同一所学校呆了几个学期，所以我有很多关于同一名学生和同一所学校的参赛作品（只有学期发生变化，即 EnrollmentBegin 和 EnrollmentEnd）。

FirstName LastName CollegeName State PublicPrivate EnrollmentBegin EnrollmentEnd
John      Doe    School A      NY       Public         20050829      20051223
John      Doe    School A      NY       Public         20051229      20060113
John      Doe    School A      NY       Public         20051223      20060513
John      Doe    School B      IL       Private        20090105      20090301
John      Doe    School B      IL       Private        20090706      20090830
John      Doe    School B      IL       Private        20090831      20091025
Jane      Doe    School A      IL       Private        20100105      20100301
Jane      Doe    School A      IL       Private        20100706      20100830
Jane      Doe    School A      IL       Private        20100831      20101025
John      Doe    School A      NY       Public         20110829      20111223
John      Doe    School A      NY       Public         20120129      20120513

这意味着对于一些学生，我有很多学生姓名和学院名称相同的条目。

我真的只想要每个新条目的第一个实例（即每当给定学生的学校名称更改时）但我还需要知道学生在那所学校的入学时间何时结束。

此信息可在给定学校每个学生的最后一个条目中找到。所以我需要从最后一个条目中获取该值，并添加到包含学生第一个条目的行中的新列。

注意：我意识到有些学生，比如上面的 John Doe，去学校 A，去另一所学校，然后回到学校 A。理想情况下，捕捉那，我希望我的最终数据集看起来像这样：

FirstName LastName CollegeName State PublicPrivate EnrollmentBegin EnrollmentEnd EnrollmentEnd
John      Doe    School A      NY       Public      20050829      20051223      20060513
John      Doe    School A      NY       Public      20110829      20111223      20120513
John      Doe    School B      IL       Private     20090105      20090301      20091025
Jane      Doe    School A      IL       Private     20100105      20100301      20101025

如何以最有效的方式执行此操作？好像min和max已经不能解决这个问题了...

Answer 1

尝试

library(data.table)
setDT(df1)[,list(EnrollmentBegin= EnrollmentBegin[1L], 
                   EnrollmentEnd=EnrollmentEnd[1L], 
                   EnrollmentEnd2= EnrollmentEnd[.N]) ,
   by =c(names(df1)[1:5])]
#   FirstName LastName CollegeName State PublicPrivate EnrollmentBegin
#1:      John      Doe    School A    NY        Public        20050829
#2:      John      Doe    School B    IL       Private        20090105
#3:      Jane      Doe    School A    IL       Private        20100105
#   EnrollmentEnd EnrollmentEnd2
#1:      20051223       20060513
#2:      20090301       20091025
#3:      20100301       20101025

或使用dplyr

library(dplyr)
df1 %>%
  group_by_(.dots=names(df1)[1:5]) %>% 
  summarise(EnrollmentBegin=EnrollmentBegin[1L], 
            EnrollmentEnd1=EnrollmentEnd[1L],
            EnrollmentEnd2 = EnrollmentEnd[n()])

Answer 2

使用基数 R 的替代方案 lapply

lst = unname(split(dat, dat[,1:5])[lapply(split(dat, dat[,1:5]), nrow) != 0])
out = do.call(rbind, lapply(lst, 
              function(x){x$EnrollmentEnd.new = x$EnrollmentEnd[nrow(x)]; x[1,]}))

#> out
#  FirstName LastName CollegeName State PublicPrivate EnrollmentBegin
#7      Jane      Doe    School_A    IL       Private        20100105
#4      John      Doe    School_B    IL       Private        20090105
#3      John      Doe    School_A    NY        Public        20050829
#  EnrollmentEnd EnrollmentEnd.new
#7      20100301          20101025
#4      20090301          20091025
#3      20051223          20060513

删除行 w。重复信息，同时保留第一个非重复条目（并将重复条目中的数据附加到该行）

Deleting rows w. duplicate information while keeping first non-duplicated entry (and appending that row with data from duplicate entries)

r

duplicate-removal

dataframe