如何基于Allele与Match的唯一配对筛选R数据框子集？

Hey there! Let's figure out why your code isn't working and get you the right result for filtering rows where the Allele + Match pair only appears once.

First, let's formalize your data into a proper data frame (I assume you might have skipped this step in your snippet):

Value <- c(FALSE, TRUE, FALSE, TRUE, FALSE)
Allele <- c('a','a','a','b','b')
Match <- c('b','b','c','b','b')
x <- data.frame(Value, Allele, Match)

What's wrong with your original code? #

The duplicated() function operates on data objects (like columns of a data frame), but you passed it two string literals: "Allele" and "Match". This treats those two strings as a short vector with no duplicates, so duplicated() returns c(FALSE, FALSE). When you use ! on that and subset your data frame, R recycles the TRUE values to match all rows—hence you get the full original data back.

Fixes that work #

Option 1: Base R with duplicated() (check both directions)

To check for duplicate pairs of Allele and Match, you need to pass those two columns as a single object to duplicated(). Also, since duplicated() only marks the second and later occurrences of a duplicate, we need to check from the end too to catch the first occurrence of a repeated pair:

# Identify all rows where the Allele-Match pair is duplicated (in either direction)
duplicate_rows <- duplicated(x[, c("Allele", "Match")]) | duplicated(x[, c("Allele", "Match")], fromLast = TRUE)

# Keep only rows that are NOT part of a duplicated pair
filtered_x <- x[!duplicate_rows, ]

Running this gives you the only row where the pair is unique:

Value Allele Match
3 FALSE      a     c

(Quick note: You mentioned expecting FALSE,a,b, but in your data, the a-b pair appears twice (rows 1 and 2), so it's actually not unique. The truly unique pair is a-c in row 3!)

Option 2: Tidyverse/dplyr approach (more readable)

If you prefer using the tidyverse, grouping by the pair and filtering for groups with only one row is super intuitive:

library(dplyr)

filtered_x <- x %>%
  group_by(Allele, Match) %>%
  filter(n() == 1) %>%
  ungroup()

This does the exact same thing as the base R method, and the code reads like plain English: group by the two columns, keep groups where the number of rows (n()) is 1, then remove the grouping.

内容的提问来源于stack exchange，提问作者Ascaris