Table and group_by in dplyr

Question

a<-data_frame(id=c(rep("A",100),rep("B",78),rep("C",50)),
              vall=c(rep("x",100),rep("y",78),rep("x",50)),
              val=runif(228))

Now, I´d like to have the following output:

a2<-a %>% group_by(id) %>% count(vall)
table(a2$vall)

but, I´d prefer to have a oneliner and not having to create a2. Is there a smart way of doing this?

Answer 1

This gets the output you want without creating a new table or using the table function. However the output layout is slightly different:

library(tidyverse)
a %>% group_by(id) %>% count(vall) %>% group_by(vall) %>% count()

Does that help at all?

Answer 2

Here is an one-liner with data.table

library(data.table)
setDT(a)[, .N, .(id, vall)][, .N , vall]

Answer 3

Akruns answer is fine, but the question was for dplyr. I just pasted the alternatives from the comments section here for completeness.

a %>% count(vall, id) %>% {table(.$vall)}

count(a, vall, id) %>% count(vall)