R语言：如何用dplyr筛选含父母子女的多成员家庭组？

Hey Julia, sorry to hear you're hitting snags with filtering your family participant data—let's get this sorted out step by step.

First, let's align on your core requirements: you want to retain only families that meet two rules:

• The family has at least 2 members
• The family includes at least one parent AND at least one child (so family_number 2, which only has parents, gets dropped)

Common Pitfalls You Might Have Hit #

From your description, it sounds like you might have used incorrect logical checks in your filter() call. For example, using all(role == "parent") && all(role == "child") would never work (a group can't be all parents and all children at the same time), or maybe you forgot to properly count the presence of each role type in the group.

Working Solution #

Assuming your dataset has columns like family_number (unique ID for each family) and role (with values like "parent" or "child"), here's a reliable approach with dplyr:

library(dplyr)

# Replace `family_data` with your actual dataset name
filtered_families <- family_data %>%
  group_by(family_number) %>%
  filter(
    # Condition 1: At least 2 members in the family
    n() >= 2,
    # Condition 2: At least one parent exists in the family
    sum(role == "parent") >= 1,
    # Condition 3: At least one child exists in the family
    sum(role == "child") >= 1
  ) %>%
  # Optional: Ungroup if you don't need family grouping for later steps
  ungroup()

Breaking Down the Logic #

• n(): When grouped by family_number, this returns the total number of members in each family—perfect for validating the minimum size requirement.
• sum(role == "parent"): Counts how many members in the family have the "parent" role. Using >= 1 ensures there's at least one parent present.
• sum(role == "child"): Does the same for children, guaranteeing at least one child is part of the family.

Handling Edge Cases #

If your role column uses more specific labels (like "mother", "father", "son", "daughter"), adjust the sum checks to include all relevant values:

filter(
  n() >= 2,
  sum(role %in% c("mother", "father")) >= 1,
  sum(role %in% c("son", "daughter")) >= 1
)

Also, watch out for case sensitivity! If some entries are "Parent" instead of "parent", standardize the values first:

family_data <- family_data %>%
  mutate(role = tolower(role))

Test with Sample Data #

Let's verify with a quick example:

# Sample dataset
family_data <- tibble(
  family_number = c(1, 1, 2, 3, 3, 3),
  role = c("parent", "child", "parent", "parent", "parent", "child")
)

# Apply the filter
filtered_families <- family_data %>%
  group_by(family_number) %>%
  filter(n() >=2, sum(role == "parent") >=1, sum(role == "child") >=1) %>%
  ungroup()

This will keep families 1 and 3, and drop family 2—exactly what you need.

内容的提问来源于stack exchange，提问作者Julia M