根据R中其余的唯一性汇总一列

Question

I want to aggregate the freq column based on the unique of the rest of the columns. I usually use

 aggregate(freq~var1+var2+var3, df, sum)

but in this case, the variables would be unknown as the user would select variables using a drop down in shiny

DISEASE Gender  Age Race    Date       Freq
Campy       F   32  ASIAN   1/6/2014    10
Campy       M   52  WHITE   1/2/2014    15
Campy       M   63  WHITE   1/3/2014    56
Chlamydia   F   24  LATINO  1/8/2014    32
Chlamydia   F   32  AA      1/3/2014    52

So in my case the user would select either of the first four columns, for instance, someone may just select the Disease, then I need to aggregate frequencies by disease and someone may select disease and gender and likewise

Answer 1

User supplies column names to aggregate as a vector:

user.cols <- c("DISEASE",'Age')

You create a unique vector in the dataframe by these columns

df$unq <- apply( df[ , user.cols ] , 1 , paste , collapse = "-" )

go ahead and aggregate by the unique column:

output <- aggregate(freq~unq, df, sum)

Answer 2

Consider a dynamic formula built with paste(..., collapse=...) :

dynamic_agg <- function(vars) {
  f <- as.formula(paste0("Freq~", paste(vars, collapse="+")))
  aggregate(f, df, FUN=sum)
}  

dynamic_agg(c("Date"))
#       Date Freq
# 1 1/2/2014   15
# 2 1/3/2014  108
# 3 1/6/2014   10
# 4 1/8/2014   32

dynamic_agg(c("Gender", "Date"))
#   Gender     Date Freq
# 1      M 1/2/2014   15
# 2      F 1/3/2014   52
# 3      M 1/3/2014   56
# 4      F 1/6/2014   10
# 5      F 1/8/2014   32

dynamic_agg(c("DISEASE", "Gender", "Date"))
#     DISEASE Gender     Date Freq
# 1     Campy      M 1/2/2014   15
# 2 Chlamydia      F 1/3/2014   52
# 3     Campy      M 1/3/2014   56
# 4     Campy      F 1/6/2014   10
# 5 Chlamydia      F 1/8/2014   32