Aggregating a dataframe dynamically with column names in strings from another dataframe

Question

Ok, so I have a configuration dataframe that will specify, how I want to aggregate a few columns in another dataframe.

config frame:

conf <- rbind(            


        

Answer 1

This looks easier to me (after correcting your sinful ways):

dd <- data.frame( c("a1", "a1", "a1"), c("b1", "b1", "b2"), c(1, 2, 3))
names(dd) <- c("a", "b", "cnt")

aggregate(dd$cnt , dd[ as.character(conf[, 1]) ], sum)
#--------
   a  b x
1 a1 b1 3
2 a1 b2 3

Answer 2

Stealing @Justin's suggestion here's what I ended up doing.

aggregate(as.formula(paste("cnt", 
                           paste(conf[,1], collapse=" + "), 
                           sep=" ~ ")), data=dd, sum)