1

我有一个看起来像这样的数据框:

w<-read.table(header=TRUE, text="
start.date   end.date   manager
2006-05-01   2007-04-30 a
2006-09-30   2007-12-31 b
1999-09-30   2007-12-31 c
2008-01-01   2012-04-30 d
2008-01-01   2020-02-28 e
2009-05-01   2016-04-08 f")

我想获得一个数据框,该数据框返回该期间每个月哪些经理正在工作,例如

df<-read.table(header=TRUE, text="
month    manager1  manager2  manager3  manager4
01-2006  a         b         c         NA
02-2006  a         b         c         d
03-2006  b         c         d         NA
04-2006  b         d         NA        NA")

我首先定义了一个函数datseq,它返回start.dateend.date之间的月份

datseq <- function(t1, t2) { 
  format(seq.Date(from = as.Date(t1,"%Y-%m-%d"), 
             to = as.Date(t2,"%Y-%m-%d"),by="month"), 
         "%m/%Y") 

但是我无法创建适当的循环来获得所需的结果。先谢谢大家的回复!

4

1 回答 1

1

由于您只需要知道月级别的重叠而不是日级别的重叠,因此您可以认为经理从第 1 天开始并在月的最后一天离开。这可以使用floor_dateceiling_datefrom package来实现lubridate

library(lubridate)

w.extended <- w

w.extended$start.date <- floor_date(as.Date(w.extended$start.date), "month")
w.extended$end.date <- ceiling_date(as.Date(w.extended$end.date), "month") - 1

#List of months
timeperiod <- seq(min(w.extended$start.date),
                  by = "month", 
                  to = max(w.extended$end.date))

然后,您可以a %within% b从包中使用lubridate它可以检查日期是否在间隔列表内。使用您提供的间隔将此功能应用于您的每个月。

df <- data.frame(t(sapply(timeperiod, 
                          function(x){
                            managersWorking <- x %within% interval(w.extended$start.date, 
                                                                   w.extended$end.date)
                            c(as.character(x), managersWorking)
                            })),
                 stringsAsFactors = F)

#Replace the 'character' format of columns to the appropriate one
df[-1] <- apply(df[-1], 2, as.logical)
df[,1]<- format(as.Date(df[,1]), "%Y/%m")

colnames(df) <- c("month", paste0("manager.", w$manager))

head(df)
#    month manager.a manager.b manager.c manager.d manager.e manager.f
#1 1999/09     FALSE     FALSE      TRUE     FALSE     FALSE     FALSE
#2 1999/10     FALSE     FALSE      TRUE     FALSE     FALSE     FALSE
#3 1999/11     FALSE     FALSE      TRUE     FALSE     FALSE     FALSE
#4 1999/12     FALSE     FALSE      TRUE     FALSE     FALSE     FALSE
#5 2000/01     FALSE     FALSE      TRUE     FALSE     FALSE     FALSE
#6 2000/02     FALSE     FALSE      TRUE     FALSE     FALSE     FALSE

原始数据:

w <- read.table(header=TRUE, text="
start.date   end.date   manager
2006-05-01   2007-04-30 a
2006-09-30   2007-12-31 b
1999-09-30   2007-12-31 c
2008-01-01   2012-04-30 d
2008-01-01   2020-02-28 e
2009-05-01   2016-04-08 f")
w
#  start.date   end.date manager
#1 2006-05-01 2007-04-30       a
#2 2006-09-30 2007-12-31       b
#3 1999-09-30 2007-12-31       c
#4 2008-01-01 2012-04-30       d
#5 2008-01-01 2020-02-28       e
#6 2009-05-01 2016-04-08       f
于 2021-03-26T11:58:59.580 回答