Get mean per observation based in characteristics in another dataframe in r

We can use base R split to split the second dataset based on the matching values of 'var1' from the first data, Filter the list elements to apply the rowMeans only if there are more than 1 column, then loop over the list, apply the rowMeans and assign those as new columns in 'df2'

lst1 <- split.default(df2, df1$var2[match(names(df2), df1$var1)])
lst2 <- Filter(function(x) ncol(x) > 1, lst1)
tmp <-  lapply(lst2,
      rowMeans, na.rm = TRUE)
df2[names(tmp)] <- tmp

-output

> df2
  X01 X02 X03  Red
1   1 0.1 2.1 1.55
2   2 0.2 2.2 2.10
3   3 0.3 2.3 2.65
4   4 0.4 2.4 3.20
5   5 0.5 2.5 3.75
6   6 0.6 2.6 4.30

Or in a for loop

for(nm in unique(df1$var2)) {
   v1 <- intersect(df1$var1[df1$var2 == nm], names(df2))
   if(length(v1) > 1) df2[[nm]] <- rowMeans(df2[v1], na.rm = TRUE)
}

data

df1 <- structure(list(var1 = c("X01", "X02", "X03", "X04", "X05", "X06"
), var2 = c("Red", "Green", "Red", "Yellow", "Red", "Green")), 
class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6"))

df2 <- structure(list(X01 = 1:6, X02 = c(0.1, 0.2, 0.3, 0.4, 0.5, 0.6
), X03 = c(2.1, 2.2, 2.3, 2.4, 2.5, 2.6)), 
class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6"))

library(tidyverse)

df2 %>%
  rowid_to_column() %>%
  pivot_longer(-rowid, names_to = 'var1') %>%
  left_join(df1) %>%
  pivot_wider(rowid, names_from = var2, values_fn = mean) %>%
  cbind(df2,.)

  X01 X02 X03 rowid  Red Green
1   1 0.1 2.1     1 1.55   0.1
2   2 0.2 2.2     2 2.10   0.2
3   3 0.3 2.3     3 2.65   0.3
4   4 0.4 2.4     4 3.20   0.4
5   5 0.5 2.5     5 3.75   0.5
6   6 0.6 2.6     6 4.30   0.6