R-人力资源中的编码

Question

我有一个数据框架，如下所示。

ID号是人员的ID，职位是他们的工作职位，Dt_Alter是他们转换角色的日期，学历是他们的背景。

我需要计算他们在岗位上的平均时间(直到他们进入经理岗位)，以及他们在进入经理岗位之前变换了多少次角色。

任何技巧我都很欣赏，因为我是R的新手，在这一部分的分析中遇到了困难。数据帧是巨大的。

ID Number        Position       Dt_Alter    Education
        2         MANAGER     2019-02-01    BUSINESS, MANAGEMENT AND ADMINISTRATION
        2     COORDINATOR     2019-01-01    BUSINESS, MANAGEMENT AND ADMINISTRATION
  2000261         MANAGER     2018-12-01    BUSINESS, MANAGEMENT AND ADMINISTRATION
  2000261      SUPERVISOR     2016-12-01    BUSINESS, MANAGEMENT AND ADMINISTRATION
  2000553         MANAGER     2018-12-01    ENGINEERING
  2000553     COORDINATOR     2016-04-01    ENGINEERING




structure(list(Matricula = c(2L, 2L, 2L, 2L, 2L), 
Desc2 = c("GERENTE", "COORDENADOR SEGUROS", "COORDENADOR SEGUROS", "COORDENADOR SEGUROS", "COORDENADOR SEGUROS"), 
Dt_Alteracao = c("01/02/2019", "01/01/2019", "01/01/2018", "01/09/2017", "01/09/2016"), 
Education = c("BUSINESS, MANAGEMENT AND ADMINISTRATION", "BUSINESS, MANAGEMENT AND ADMINISTRATION", "BUSINESS, MANAGEMENT AND ADMINISTRATION",  "BUSINESS, MANAGEMENT AND ADMINISTRATION", "BUSINESS, MANAGEMENT AND ADMINISTRATION")), 
row.names = c("2.10823", "2.10824", "2.10825", "2.10826", "2.10827"), class = "data.frame")

Answer 1

下面是我结合使用ifelse和lag函数来解决这个问题的粗略方法。基本上，在确保您已经在ID.Number和date上对文件进行排序之后，记录的顺序应该允许跨记录进行比较。我为某人是否改变了他们的位置做了一个标记，如果是真的，就计算这些记录的difftime。

希望这能有所帮助。

df$Matricula<-as.character(df$Matricula)
df$Dt_Alteracao<-strptime(df$Dt_Alteracao,format="%d/%m/%Y")
df<-df[order(df$Matricula, df$Dt_Alteracao), ]

# indicator for whether a position change occurred
df$changePos<-ifelse( df$Matricula== lag(df$Matricula,1) & df$Desc2 != lag(df$Desc2,1), 
                  "Changed Position", "Same")   # review this logic for a variety of row groupings

# measure weeks between positions
library(lubridate)
df$Dt_Alteracao2<-as.POSIXct(df$Dt_Alteracao)
df$time_in_pos<-ifelse(df$changePos=="Changed Position", 
                   difftime(lag(df$Dt_Alteracao2,1),df$Dt_Alteracao2,units ='weeks'),NA )