如何基于不同分组条件筛选R语言中的DataFrame？

Hey there! Let's work through this conditional filtering problem step by step. Since you need different MG selection rules based on specific ranges of loc.id, dplyr's case_when() function is perfect here—it lets you define clear, scenario-based conditions directly inside filter().

Step 1: Set up your data (with reproducibility) #

First, let's recreate your data frame with a seed so the random x values are consistent every time you run it:

library(dplyr)

set.seed(123) # Ensures runif() gives the same values each time
df <- data.frame(
  loc.id = rep(1:10, each = 10),
  MG = rep(1:10, times = 10),
  x = runif(100)
)

Step 2: Apply the conditional filter #

Here's the code to implement your exact requirements:

filtered_df <- df %>%
  filter(
    case_when(
      # When loc.id is less than 4: keep MG 1-4
      loc.id < 4 ~ MG %in% 1:4,
      # When loc.id is between 5 and 6 (inclusive): keep MG 5-8
      between(loc.id, 5, 6) ~ MG %in% 5:8,
      # When loc.id is greater than 6: keep MG >8 (i.e., 9-10)
      loc.id > 6 ~ MG > 8,
      # Catch-all: exclude any rows that don't match the above (optional here)
      TRUE ~ FALSE
    )
  )

How this works: #

• case_when() evaluates each row against the conditions in order. For each row, it returns TRUE or FALSE based on the first matching condition.
• MG %in% 1:4 checks if the MG value falls exactly within the 1-4 range (inclusive)—this is cleaner than writing MG >=1 & MG <=4.
• between(loc.id, 5,6) is a handy dplyr function to check if a value is within a closed range (no need for loc.id >=5 & loc.id <=6).
• The final TRUE ~ FALSE line is a safety net to exclude any rows that don't fit your defined loc.id ranges (though in your data, loc.id only goes 1-10, so this is optional, but good practice for robustness).

Verify the results #

To make sure the filtering worked as expected, you can check the unique combinations of loc.id and MG in the filtered data:

filtered_df %>%
  distinct(loc.id, MG) %>%
  arrange(loc.id, MG)

This will show you exactly which pairs are kept, confirming that:

• loc.id 1-3 only have MG 1-4
• loc.id 5-6 only have MG 5-8
• loc.id7-10 only have MG 9-10

内容的提问来源于stack exchange，提问作者89_Simple