R语言多列条件合并/替换求助（基于唯一列组合）

Got it, since your key columns (x1 and x2) have unique combinations, this makes the conditional replacement straightforward. Below are two common, practical approaches in R—one using the tidyverse (dplyr) and another using data.table (ideal for large datasets):

First, let's recap your sample data for clarity:

df1 <- data.frame(x1 = c(1,1,2,2), x2 = rep(letters[1:2],2), x3 = letters[1:4], stringsAsFactors = FALSE )
df2 <- data.frame(x1 = c(1,2,3), x2 = letters[1:3], x3 = c('xx','zz','yy'), stringsAsFactors = FALSE )

Using dplyr (tidyverse) #

This method uses a left join to pull in values from df2, then replaces df1's x3 with df2's where matches exist:

library(dplyr)

df_updated <- df1 %>%
  left_join(df2, by = c("x1", "x2"), suffix = c("_original", "_new")) %>%
  mutate(x3 = coalesce(x3_new, x3_original)) %>%
  select(x1, x2, x3)

# View the final result
print(df_updated)

Breaking this down:

• left_join keeps all rows from df1 and adds matching rows from df2, using x1 and x2 as the join keys. The suffixes help tell the original x3 apart from the new one.
• coalesce picks the first non-NA value—so it uses df2's x3 if available, otherwise sticks with df1's original value.
• select cleans up the columns to get back your original table structure.

Using data.table (efficient for large datasets) #

If you're working with big data, data.table's update join is faster and uses less memory because it modifies data in place (or creates a copy if you prefer):

library(data.table)

# Convert data frames to data.table objects
setDT(df1)
setDT(df2)

# Perform update join: replace x3 in df1 with df2's x3 where x1 and x2 match
df1[df2, x3 := i.x3, on = c("x1", "x2")]

# View the updated df1
print(df1)

Here's what's happening:

• setDT converts your standard data frames to data.table objects for efficient operations.
• The syntax df1[df2, x3 := i.x3, on = c("x1", "x2")] means: for each row in df2, find the matching row in df1 using x1 and x2, then set df1's x3 to df2's x3 (denoted by i.x3, where i refers to the "lookup" table—df2 here). Rows in df1 without a match stay unchanged.

Both methods will give you the same result: rows where x1 and x2 match between df1 and df2 get their x3 updated to df2's value, while non-matching rows retain their original x3.

内容的提问来源于stack exchange，提问作者tjebo