group_by、do()和更多链/多个条件之后dplyr内的if语句

Question

我希望你们都很好。

我有一个包含许多列的数据集，我正在尝试根据多个条件删除重复的数据集。下面我提供一个示例来演示我的问题。其思想是，对于每个ID，检查所有列，如果所有列都相同，则保留最新的列。如果有两个相同的行，而上面的注释是不同的，那么检查该行是否为"Add comment for down/upgrading client"，如果所有行都有相同的注释，则保留第一行，否则保留最新的行，不包含上面的注释。

我一直在尝试以下几种方法

##dataframe
             ID <- c("H1", "H1"," H1"," H2", "H2", "H3", "H3"," H3", "H4")
            rating <-c("C", "C", "C+","D", "C", "C",  "C+", "C+", "C")
            Commnets<- c("Add comment for down/upgrading client", "updated", "Add comment for down/upgrading client","Add comment for down/upgrading client","Add comment for down/upgrading client", 
                        "down",  "down", "Add comment for down/upgrading client", "Add comment for down/upgrading client")
            Date<- c("2018-12-10", "2018-12-10", "2018-11-10",
                        "2018-11-10","2018-11-10", 
                        "2018-10-10",  "2018-10-02", "2018-10-02", "2020-09-03")
 df<-data.frame(ID,rating,Commnets,Date,stringsAsFactors=FALSE)






 df$Date<-as.Date(df$Date)
    df<-df%>%
      group_by(ID,rating,Date)%>%
      arrange(desc(Date)) %>% # in each group, arrange in desc by Date
      filter(row_number() == 1)#this will solve the first problem 



  
   



df$Date<-as.Date(df$Date)
        df<-df%>%
          group_by(ID,rating,Date)%>%
          arrange(desc(Date)) %>% #I think that I need **do** here but not sure how
ifelse(rowSums("Add comment for down/upgrading client" == $Comments)==length($Comments),
                  filter(row_number() == 1),rowSums("Add comment for down/upgrading client" == $Comments)[1,])

Answer 1

您可以通过递减Date顺序和计数每个ID、rating和Date的唯一Commnets数来arrange数据。如果始终都是相同的注释，则选择第一行，如果不同，则选择最后一行，即最新的。

library(dplyr)

df %>%
  mutate(ID = trimws(ID), 
         Date = as.Date(Date)) %>%
  arrange(ID, rating, Commnets, desc(Date)) %>%
  group_by(ID,rating,Date)  %>%
  slice(if(n_distinct(Commnets) == 1) 1L else n())

#  ID    rating Commnets                              Date      
#  <chr> <chr>  <chr>                                 <date>    
#1 H1    C      updated                               2018-12-10
#2 H1    C+     Add comment for down/upgrading client 2018-11-10
#3 H2    C      Add comment for down/upgrading client 2018-11-10
#4 H2    D      Add comment for down/upgrading client 2018-11-10
#5 H3    C      down                                  2018-10-10
#6 H3    C+     down                                  2018-10-02
#7 H4    C      Add comment for down/upgrading client 2020-09-03