updating column in one dataframe with value from another dataframe based on matching values

you don't need z$color in the first place if its just place holder, you can replace NA later with 0

z$color<-y[match(z$letter, y$letter),2]

You can use merge:

dat <- merge(z, y, by = "letter", all.x = TRUE)
transform(dat, color = ifelse(is.na(color.y), 
                              color.x, as.character(color.y)))[-(2:3)]

   letter color
1       a   red
2       a   red
3       a   red
4       b  blue
5       b  blue
6       b  blue
7       b  blue
8       c green
9       c green
10      c green
11      c green
12      c green
13      d     0
14      d     0
15      d     0
16      d     0
17      e     0
18      e     0
19      e     0
20      e     0

sqldf/sqlite is very flexible:

library(sqldf)
z$color="0" # to avoid conflicts numeric/characters
z <- sqldf(c("UPDATE z
             SET color = (SELECT y.color
                          FROM y
                          WHERE z.letter = y.letter
                           )
             WHERE EXISTS (SELECT 1
                           FROM y
                           WHERE z.letter = y.letter
                           )"
             , "select * from main.z"
                  )
           )
z
   letter color
1       b  blue
2       a   red
3       d   0.0
4       d   0.0
5       e   0.0
6       a   red
7       a   red
8       c green
9       b  blue
10      c green
11      e   0.0
12      c green
13      b  blue
14      d   0.0
15      d   0.0
16      d   0.0
17      c green
18      e   0.0
19      a   red
20      c green