删除重复项， dplyr 删除非重复行。

Question

我正在尝试向数据框中添加行，然后检查/删除在数据框的单个列中具有重复值的行。最终目标是如果提供了新值，则本质上是覆盖数据框中的一行。我无法弄清楚如何使用 dplyr （或实际上使用 R）动态指定行名，所以我正在这样做。

我从一个测试数据框开始，并使用 dplyr 删除列位置重复的第一组行，如下所示：

    testData.df<-data_frame(Position=c("B1","B2","B3","B1","B2","B3"), rep=c("B1","B2","B3","B4","B5","B6"),name=c(rep("wibble",each=6)),status=c(rep("unknown", each=6)))
    testData.df <- testData.df %>%
      filter(duplicated(Position))
    testData.df
# A tibble: 3 x 4
  Position   rep   name  status
     <chr> <chr>  <chr>   <chr>
1       B1    B4 wibble unknown
2       B2    B5 wibble unknown
3       B3    B6 wibble unknown

正如我所期望的那样。当我再次运行相同的过滤器时，我得到了这个：

testData.df <- testData.df %>%
  filter(duplicated(Position))
testData.df
# A tibble: 0 x 4
# ... with 4 variables: Position <chr>, rep <chr>, name <chr>, status <chr>

到底为什么要删除不重复的行？第一次运行它意味着它正在按预期工作，即它删除了实际的重复项。我无法解释第二轮的行为差异。

Answer 1

您期望filter(duplicated(...)) 保留非重复行，但实际上却相反。如果我们将row_numbers 添加到每一行，您可以看到这一点

testData.df<-data_frame(Position=c("B1","B2","B3","B1","B2","B3"), rep=c("B1","B2","B3","B4","B5","B6"),name=c(rep("wibble",each=6)),status=c(rep("unknown", each=6))) %>%
               mutate(rn = row_number())
testData.df <- testData.df %>%
                  filter(duplicated(Position))

输出

# A tibble: 3 x 5
  Position   rep   name  status    rn
     <chr> <chr>  <chr>   <chr> <int>
1       B1    B4 wibble unknown     4
2       B2    B5 wibble unknown     5
3       B3    B6 wibble unknown     6

你应该改用filter(!duplicated(...))

---

已编辑

第一次尝试保留重复的行，但第二次不会丢失它

testData.df<-data_frame(Position=c("B1","B2","B3","B1","B2","B3"), rep=c("B1","B2","B3","B4","B5","B6"),name=c(rep("wibble",each=6)),status=c(rep("unknown", each=6))) %>%
                mutate(rn = row_number())

run1 <- testData.df %>%
        group_by(Position) %>%
        slice(n()) %>%
            ungroup()

run2 <- run1 %>%
        group_by(Position) %>%
        slice(n()) %>%
            ungroup()

# A tibble: 3 x 5
# Groups:   Position [3]
  Position   rep   name  status    rn
     <chr> <chr>  <chr>   <chr> <int>
1       B1    B4 wibble unknown     4
2       B2    B5 wibble unknown     5
3       B3    B6 wibble unknown     6