Question

我有一个这样的数据框：

df <- data.frame(
  names = c(rep("cody", 10), rep("sam", 5)),
  year  = c(paste0("year",2000:2009), paste0("year",2000:2004))
)

我想获得这样的结果输出：

df2 <- data.frame(
  names = c(rep("cody", 5), rep("sam", 5)), 
  year  = c(paste0("year",2000:2004), paste0("year",2000:2004))
)

有什么想法吗？

Answer 1

以下是包含Reduce和intersect的基本R方法。

dat[dat$year == Reduce(intersect, split(dat$year, dat$names)),]

返回

  names     year
1   cody year2000
2   cody year2001
3   cody year2002
4   cody year2003
5   cody year2004
11   sam year2000
12   sam year2001
13   sam year2002
14   sam year2003
15   sam year2004

在这里，我们使用Reduce重复提供参数（使用split作为列表提供的每个名称的不同年份）到intersect，这消除了“不匹配”的年份，直到你最终只有那些可用于所有名字的年份。

请注意，年变量必须是字符向量，而不是因子变量。

作为次要简化，您可以使用with来减少dat$引用：

dat[with(dat, year == Reduce(intersect, split(year, names))),]

数据

dat <- structure(list(names = c("cody", "cody", "cody", "cody", "cody", "cody", "cody", "cody", "cody", "cody", "sam", "sam", "sam", "sam", "sam"), year = c("year2000", "year2001", "year2002", "year2003", "year2004", "year2005", "year2006", "year2007", "year2008", "year2009", "year2000", "year2001", "year2002", "year2003", "year2004")), .Names = c("names", "year"), row.names = c(NA, -15L), class = "data.frame")

Answer 2

您可以按年份进行分组，然后过滤那些出现两次的年份（或者您想要的许多唯一名称）：

library(dplyr)

df %>% 
  group_by(year) %>% 
  mutate(name_count = n()) %>%
  ungroup() %>% 
  filter(name_count == 2) %>% 
  select(-name_count)

   names year    
   <fct> <fct>   
 1 cody  year2000
 2 cody  year2001
 3 cody  year2002
 4 cody  year2003
 5 cody  year2004
 6 sam   year2000
 7 sam   year2001
 8 sam   year2002
 9 sam   year2003
10 sam   year2004

Answer 3

以下是查找year列中所有重复项的选项。

df[duplicated(df$year) | duplicated(df$year, fromLast = TRUE), ]
#    names     year
# 1   cody year2000
# 2   cody year2001
# 3   cody year2002
# 4   cody year2003
# 5   cody year2004
# 11   sam year2000
# 12   sam year2001
# 13   sam year2002
# 14   sam year2003
# 15   sam year2004

在R中查找数据帧的分组变量的交集

3 个答案: