Conditional mean from different column

I do have an R data frame like this:

city2001 <- c('a', 'b', 'a')
grade2001 <- c(5, 5, 7)
city2002 <- c('b', 'b', 'a')
grade2002 <- c(8, 9, 10)

df <- data.frame(city2001, grade2001, city2002, grade2002)

and would like to return ,

avg_a = 7.333
# from (5 + 7 + 10)/3

How is the logic for that? Thanks.

>Solution :

Try

mean(df[,grepl("grade",colnames(df))][df[,grepl("city",colnames(df))]=="a"])
[1] 7.333333

your df (columns) better be sorted.

If you want for all the groups and not just "a"

tapply(
  unlist(df[,grepl("grade",colnames(df))]),
  unlist(df[,grepl("city",colnames(df))]),
  mean
)
       a        b 
7.333333 7.333333