如何在R语言中基于另外两列的特定行值创建新变量？

Got it, let's work through this problem step by step. First, I'll assume your dataset has a structure like this (since each patient has 4 rows, it's almost certainly split by ear and time point):

Your goal is to create a new variable that pulls chemo_dose from:

• The first row of "chemo dose 1" per patient (likely the left ear row)
• The second row of "chemo dose 2" per patient (likely the right ear row)

Here are two reliable ways to implement this in R:

Method 1: Using dplyr (Tidyverse) #

This is the most readable approach, especially if you're already working with tidyverse tools. We'll group by patient and target the exact rows we need.

Option A: Target rows by ear (most reliable) #

If you have an ear column marking left/right, use this instead of relying on row order (which can shift unexpectedly):

library(dplyr)

df_new <- df %>%
  group_by(patient_id) %>%
  mutate(
    new_chemo_var = case_when(
      # Pull value from left ear during chemo dose 1
      time_point == "chemo dose 1" & ear == "L" ~ chemo_dose,
      # Pull value from right ear during chemo dose 2
      time_point == "chemo dose 2" & ear == "R" ~ chemo_dose,
      # Set all other rows to NA (adjust this if you need a different default)
      TRUE ~ NA_real_
    )
  ) %>%
  ungroup()

Option B: Target rows by position (if row order is guaranteed) #

If you're 100% sure each patient's rows follow the order: left dose1 → right dose1 → left dose2 → right dose2, you can use row numbers within each patient group:

df_new <- df %>%
  group_by(patient_id) %>%
  mutate(
    new_chemo_var = case_when(
      # 1st row in group = left ear dose1
      row_number() == 1 ~ chemo_dose,
      # 4th row in group = right ear dose2
      row_number() == 4 ~ chemo_dose,
      TRUE ~ NA_real_
    )
  ) %>%
  ungroup()

Method 2: Using Base R #

If you prefer not to load additional libraries, here's a base R approach using splitting and looping:

# Initialize the new variable with NA values
df$new_chemo_var <- NA_real_

# Split the dataset into patient-specific groups
patient_groups <- split(df, df$patient_id)

# Iterate over each patient group to assign values
for (group_name in names(patient_groups)) {
  group <- patient_groups[[group_name]]
  
  # Find indices for the rows we need
  dose1_left_idx <- which(group$time_point == "chemo dose 1" & group$ear == "L")
  dose2_right_idx <- which(group$time_point == "chemo dose 2" & group$ear == "R")
  
  # Assign the chemo dose values to the new variable
  if (length(dose1_left_idx) > 0) {
    group$new_chemo_var[dose1_left_idx] <- group$chemo_dose[dose1_left_idx]
  }
  if (length(dose2_right_idx) > 0) {
    group$new_chemo_var[dose2_right_idx] <- group$chemo_dose[dose2_right_idx]
  }
  
  # Update the group in the list
  patient_groups[[group_name]] <- group
}

# Combine all groups back into a single data frame
df_new <- do.call(rbind, patient_groups)

Quick Tips #

• Always prioritize using explicit identifiers like ear over row order—row order can change during data cleaning or importing, leading to errors.
• If you want to propagate these values to all rows for each patient (instead of just the target rows), you can use first()/last() within the group to pull the value and assign it to every row.

内容的提问来源于stack exchange，提问作者sabc04