不带NA的连接字符串

Question

我想总结一下没有na的字符串。

它正在使用str_c，如下面的示例所示

码

df <- tibble(x = c("ISSUE-1", "ISSUE-2", "ISSUE-3", "ISSUE-4"), ID = c("ID-1", "ID-1", "ID-2", "ID-2")) %>%
  group_by(ID) %>%
  summarise(IDissue = str_c(x, collapse = "; "))

输出

ID          issueID
ID-1    ISSUE-1; ISSUE-2
ID-2    ISSUE-3; ISSUE-4

但是，当列中有NA时，str_c将整个输出转换为NA，如文档：https://stringr.tidyverse.org/reference/str_c.html中所述。

和大多数其他R函数一样，缺少的值是“具有传染性的”：当一个缺失的值与另一个字符串组合时，结果总是会丢失的。使用str_replace_na()将NA转换为"NA“代码

df <- tibble(x = c("ISSUE-1", "ISSUE-2", "ISSUE-3", NA), ID = c("ID-1", "ID-1", "ID-2", "ID-2")) %>%
  group_by(ID) %>%
  summarise(IDissue = str_c(x, collapse = "; "))

输出

ID    IDissue
ID-1    ISSUE-1; ISSUE-2
ID-2    NA

有什么方法可以作为输出吗？

ID    IDissue
ID-1    ISSUE-1; ISSUE-2
ID-2    ISSUE-3

Answer 1

更容易在filter之前添加group_by

library(stringr)
library(dplyr)
tibble(x = c("ISSUE-1", "ISSUE-2", "ISSUE-3", NA), 
     ID = c("ID-1", "ID-1", "ID-2", "ID-2")) %>%  
  filter(!is.na(x)) %>% 
  group_by(ID) %>%
  summarise(IDissue = str_c(x, collapse = "; "))

Answer 2

基于aggregate的R基选项

> aggregate(.~ID,df,function(x) paste0(na.omit(x),collapse = "; "))
    ID                x
1 ID-1 ISSUE-1; ISSUE-2
2 ID-2          ISSUE-3

Answer 3

您可以将粘贴或str_c与x!is.na(x)一起使用，而不是x；它可以工作。

library(dplyr)    
df <- tibble(x = c("ISSUE-1", "ISSUE-2", "ISSUE-3", NA), ID = c("ID-1", "ID-1", "ID- 
          2", "ID-2")) %>%
          group_by(ID) %>%
          summarise(IDissue = paste(x[!is.na(x)], collapse = "; "))
df