r语言 - 标识共享重复值的组

Question

我有一个数据框架，总结了不同设备的部署历史，以序列号标识。随着时间的推移，设备(serno)可以被不同的项目使用，但在任何给定时间只能部署在一个项目上。我生成了一个名为used.elsewhere的新列，用于标识该行中的血清是否在df中被复制。例如:

project = c("a","b","c","c","c","d","e")
serno = c(1,2,2,2,3,3,3)
deployed = c(T,T,F,F,F,F,T)
used.elsewhere = c(F,T,T,T,T,T,T)
data.frame(project,serno,deployed,used.elsewhere)
  project serno deployed used.elsewhere
1       a     1     TRUE          FALSE
2       b     2     TRUE           TRUE
3       c     2    FALSE           TRUE
4       c     2    FALSE           TRUE
5       c     3    FALSE           TRUE
6       d     3    FALSE           TRUE
7       e     3     TRUE           TRUE

我想生成一个新列，如果没有部署serno值并且在其他地方使用，则该列指示部署serno的项目:

project = c("a","b","c","c","c","d","e")
serno = c(1,2,2,2,3,3,3)
deployed = c(T,T,F,F,F,F,T)
used.elsewhere = c(F,T,T,T,T,T,T)
other.project = c(NA, NA, "b", "b", "e", "e", NA)
  project serno deployed used.elsewhere other.project
1       a     1     TRUE          FALSE          <NA>
2       b     2     TRUE           TRUE          <NA>
3       c     2    FALSE           TRUE             b
4       c     2    FALSE           TRUE             b
5       c     3    FALSE           TRUE             e
6       d     3    FALSE           TRUE             e
7       e     3     TRUE           TRUE          <NA>

我假设我可以使用如下的ifelse语句，但我不确定如何完成它。

df %>%
  mutate(other.project = ifelse(deployed == F & used.elsewhere == T, ...

提前感谢!

Answer 1

如果您先使用group_by(serno)，则可以在同一组中包含project，其中deployed为TRUE。

library(dplyr)
df %>%
  group_by(serno) %>%
  mutate(other.project = ifelse(
    deployed == FALSE & used.elsewhere == TRUE,
    project[deployed],
    NA
  ))

  project serno deployed used.elsewhere other.project
  <chr>   <dbl> <lgl>    <lgl>          <chr>        
1 a           1 TRUE     FALSE          NA           
2 b           2 TRUE     TRUE           NA           
3 c           2 FALSE    TRUE           b            
4 c           2 FALSE    TRUE           b            
5 c           3 FALSE    TRUE           e            
6 d           3 FALSE    TRUE           e            
7 e           3 TRUE     TRUE           NA 

Answer 2

我建议构建一个部署映射，然后将它与您的df:

连接起来。

library(dplyr)
deployments <- df %>% 
  filter(deployed == TRUE) %>% 
  select(serno, other.project = project)
df %>% 
  left_join(deployments)

然后你只需要确保你在other.project列中有你想要的NAs(即，通过重新编码other.project == project)。