Use summarise_each on grouped data

Question

I am trying to summerise a grouped dataset with summerise_each. Somehow it is not possible to drop the grouping variable before summarising (I tried it with dplyr::select). The problem is that the grouping variable is not numeric and I want the sum all columns.

library(dplyr)   
group  <- sample(c("a","b","c"), 100, replace = TRUE,)
values <- matrix(rnorm(1000), ncol = 10)
data   <- data.frame(group, values)

data %>% group_by(group) %>% summarise_each(sum)

Error in UseMethod("as.lazy_dots") : no applicable method for 'as.lazy_dots' applied to an object of class "function"

I want this output but for all columns.

data %>% group_by(group) %>% summarise(sum(X1))´

 group  sum(X1)
  (fctr)    (dbl)
1      a 2.648381
2      b 5.532697
3      c 1.382693

I do not want to use summarise(sum(X2), sum(X2), ...)

akrun · Accepted Answer · 2016-06-18T14:32:54.780

We need to use the function inside funs

data %>% 
    group_by(group) %>% 
    summarise_each(funs(sum))   

#   group         X1        X2          X3        X4         X5        X6        X7          X8        X9       X10
#  <fctr>      <dbl>     <dbl>       <dbl>     <dbl>      <dbl>     <dbl>     <dbl>       <dbl>     <dbl>     <dbl>
#1      a  -1.700664  3.570311 -0.23851052 7.5914770  0.3646715 -4.547660  8.944908   0.3996188 4.4322264  3.778364
#2      b -12.535340  2.705520 -0.03063312 3.1732708 -5.8638185 -8.070278 -3.548260  -1.7781691 0.7202883 -2.634258
#3      c   9.925093 -7.477544 -0.27394536 0.8187566 -7.7432485 -6.190602 -3.785092 -10.1892073 7.9212578 -2.766949

mkearney · Answer 2 · 2016-06-21T15:15:20.270

To build a user function to complete this task, try:

summaRize <- function(x) {
  sumfun <- function(fun, x) {
    round(fun(as.numeric(as.character(x)), na.rm = TRUE), 2)
  }
  out <- data.frame(min = sumfun(min, x),
                    max = sumfun(max, x),
                    mean = sumfun(mean, x),
                    median = sumfun(median, x),
                    sd = sumfun(sd, x),
                    var = sumfun(var, x))
  out[, ] <- apply(out, 2, as.numeric)
  out
}
sapply(unique(data$group), function(x) data[data$group == x, ], summaRize)

Use summarise_each on grouped data

2 Answers2