Let’s say I have two starting data frames:

df1 <- data.frame(code1 = c("a", "b","z"), code2 = c("2", "3", "4"))
df2 <- data.frame(code1 = c("c", "o", "p"), code2 = c("2", "4", "5"), 
                  column3 = "a", column4 = "b", column5 = "c")

I want to match the two data frames by the column ‘code2’ and where that’s a match, replace the value of code1 in df1 to the value of code1 in df2 so that the final data frame looks like this:

df3<- data.frame(code1 = c("c", "b", "o"), code2 = c("2", "3", "4"))

>Solution :

Here’s a solution with dplyr. It "looks up" code1 in df2, wherever code2 matches; and when no match is found, it defaults to the original code1 in df1.

Solution

library(dplyr)


# ...
# Code to generate 'df1' and 'df2'.
# ...


df1 %>% mutate(code1 = coalesce(
  # Look up the 'code1' according to 'code2'...
  df2$code1[match(code2, df2$code2)],
  # ...and otherwise default to the original 'code1'.
  code1
))

Result

Given df1 and df2 as in your example

df1 <- data.frame(
  code1 = c("a", "b","z"),
  code2 = c("2", "3", "4")
)

df2 <- data.frame(
  code1 = c("c", "o", "p"),
  code2 = c("2", "4", "5"),
  column3 = "a",
  column4 = "b",
  column5 = "c"
)

this solution should yield the desired result:

  code1 code2
1     c     2
2     b     3
3     o     4

Note

One advantage of using match() rather than a dplyr::*_join(): no additional steps are needed to purge extraneous columns from your results.