我是R的新手,有一个名为final的数据框作为主要数据集,如下所示
dates_seq_ajay<-as.data.frame((seq(as.Date("2019/11/1"), by = "month", length.out = 6)))
ajay_emp_no <-1
ajay_ramped <-c(0,0,0,0,1,1)
ajay_loans <-c(1,22,17,25,21,23)
name<-"ajay"
data<-cbind(name,ajay_emp_no,dates_seq_ajay,ajay_ramped,ajay_loans)
colnames(data)<-c("name","emp_no","date","Flag","loans")
dates_seq_dv<-as.data.frame((seq(as.Date("2019/11/1"), by = "month", length.out = 4)))
dv_emp_no <-2
dv_flag <-c(0,0,0,0)
dv_loans <-c(2,15,42,1)
name<-"dv"
data1<-cbind(name,dv_emp_no,dates_seq_dv,dv_flag,dv_loans)
colnames(data1)<-c("name","emp_no","date","Flag","loans")
dates_seq_prince<-as.data.frame((seq(as.Date("2020/5/1"), by = "month", length.out = 5)))
prince_emp_no <-3
prince_flag <-c(0,0,0,1,1)
prince_loans <-c(16,31,28,32,23)
name<-"prince"
data2<-cbind(name,prince_emp_no,dates_seq_prince,prince_flag,prince_loans)
colnames(data2)<-c("name","emp_no","date","Flag","loans")
final<-rbind(data,data1,data2)
我的df中有1000名员工,我想为每个员工查找月数,绩效和累计绩效 这样,如果某位员工第一次遇到标志1,则可以在下面的
中根据需要进行计算如果员工标志为0并且没有标志1,则计算月,绩效和累积绩效,直到有记录为止。
对于每位员工
Month是他在场的月份数,
绩效是每月贷款/总贷款的比例
总贷款是直到第一次 找到该标志的所有贷款的总和,如果标志始终为0,则总贷款是所有贷款的总和
累积绩效是指员工在每一步中累积的贷款总额,直到我们标记1为止
输出如下所示,仅适用于3名员工,但是我需要对所有1000名员工有一个共同的逻辑
答案 0 :(得分:2)
我们按“名称”分组,通过取“标志”('tmp')的累加总和创建一个临时列,获得“ {Month”的row_number()
,然后将“ Performance”除以基于“ tmp”小于2的条件,其中sum
为“贷款”的“贷款”,“累积性能”为“性能”的总和。然后,根据情况将这些列中的行替换为NA,并以'tmp'列为条件,并删除'tmp'
library(dplyr) #1.0.0
final %>%
group_by(name) %>%
mutate(tmp = cumsum(Flag),
Month = row_number(),
Performance= loans/sum(loans[tmp <2]),
CumulativePerformance = cumsum(Performance)) %>%
mutate(across(Month:CumulativePerformance, ~ replace(., tmp > 1, NA))) %>%
ungroup %>%
select(-tmp)
# A tibble: 15 x 8
# name emp_no date Flag loans Month Performance CumulativePerformance
# <chr> <dbl> <date> <dbl> <dbl> <int> <dbl> <dbl>
# 1 ajay 1 2019-11-01 0 1 1 0.0116 0.0116
# 2 ajay 1 2019-12-01 0 22 2 0.256 0.267
# 3 ajay 1 2020-01-01 0 17 3 0.198 0.465
# 4 ajay 1 2020-02-01 0 25 4 0.291 0.756
# 5 ajay 1 2020-03-01 1 21 5 0.244 1
# 6 ajay 1 2020-04-01 1 23 NA NA NA
# 7 dv 2 2019-11-01 0 2 1 0.0333 0.0333
# 8 dv 2 2019-12-01 0 15 2 0.25 0.283
# 9 dv 2 2020-01-01 0 42 3 0.7 0.983
#10 dv 2 2020-02-01 0 1 4 0.0167 1
#11 prince 3 2020-05-01 0 16 1 0.150 0.150
#12 prince 3 2020-06-01 0 31 2 0.290 0.439
#13 prince 3 2020-07-01 0 28 3 0.262 0.701
#14 prince 3 2020-08-01 1 32 4 0.299 1.00
#15 prince 3 2020-09-01 1 23 NA NA NA
如果我们使用的是dplyr
的早期版本,请使用mutate_at
代替mutate(across
final %>%
group_by(name) %>%
mutate(tmp = cumsum(Flag),
Month = row_number(),
Performance= loans/sum(loans[tmp <2]),
CumulativePerformance = cumsum(Performance)) %>%
mutate_at(vars(Month:CumulativePerformance), ~ replace(., tmp > 1, NA)) %>%
ungroup %>%
select(-tmp)