Loop through multiple columns in R - r

Question

I have the following code I'd like to run for multiple columns in a data frame called ccc.
ccc %>%
group_by(LA) %>%
summarise(Def = sum(DefaultOct05 == 'Def'),
NDef = sum(DefaultOct05 != 'Def'),
DRate = mean(DefaultOct05 == 'Def'))
LA is the name of one of the columns. How would I set up a loop to run through a number of different columns?
I've tried the following.
for (i in 26:ncol(ccc)) {
ccc %>%
group_by(i) %>%
summarise(Def = sum(DefaultOct05 == 'Def'),
NDef = sum(DefaultOct05 != 'Def'),
DRate = mean(DefaultOct05 == 'Def'))
}
But I get the following error message.
Error in resolve_vars(new_groups, tbl_vars(.data)) :
unknown variable to group by : i

Answer 1

What most people will miss in your question is a reproducible data set. Without it, its often very hard to reproduce your problem and solve it.
If I got you right, your data-set looks like the one above:
set.seed(1)
ccc=data.frame(Default=sample(c(0,1),100,replace = TRUE),LA=sample(c("X","Y","Z"),100,replace = TRUE),DC=sample(c("A","B","C"),100,replace = TRUE))
do.call() - applies rbind() to the subsequent elements.
lapply(dat,function(x)) applies the function to every element of dat - in our case columns.
library(dplyr)
do.call(rbind,lapply(ccc, function(Var) {
dat=data.frame(Var,Default=ccc$Default) %>% group_by(Var) %>% summarise(Def=sum(Default),NDef=n()-sum(Default),DRate=mean(Default))
return(as.data.frame(dat))
}
))

Answer 2

"LA is the name of one of the columns"
Actually, group by dplyr construction works on variables inside the columns. I guess you want to do other things.
If you want to apply the same function to different columns you could use summarize_at.
df <- data.frame( id = c(1:20),
a1 = runif(20),
b1 = runif(20),
c1 = runif(20)
)
library(dplyr)
df %>% summarise_at(c("a1","b1","c1"), funs(med = median,
avr = mean))
# result:
# a1_med b1_med c1_med a1_avr b1_avr c1_avr
# 1 0.6444056 0.5266252 0.6420554 0.5605837 0.4983654 0.5546381