Counting the number of specific integers per column in an R matrix

Question

I have a matrix (10 x 100) where I need to count the number of each integer per column so I have a final matrix that is (3 x 100). Counts for 0, 1, and 2 per column.

I think the apply function will be useful here, the code I provided is a solution I envision.

Any help will be greatly appreciated.

library(dplyr)
set.seed(100)
a <- matrix(sample(0:2, size=100, replace=TRUE), nrow=10, ncol=100)
out <- apply(a, 2, function(x) count(x))

 Desired output: rows are the sum of each variable "0, 1, 2"

   1 2 3 ...  n
 0 1 1 3
 1 6 3 3
 2 3 6 4

Answer 1

There is a function called table that counts the distinct values of the whole object. You can apply it to each column, i.e.

apply(a, 2, table)

to include NAs in the count just use the option useNA, i.e.

apply(a, 2, table, useNA = 'always')

#or with complete syntax
apply(a, 2, function(i)table(i, useNA = 'always'))

As @IceCreamToucan mentions in comments, If you have missing values in any column, then you want be able to coerce to to a data frame (or matrix for that matter). To overcome this, we can convert each column to factor with levels = c(0:2), i.e.

apply(a, 2, function(i) table(factor(i, levels = c(0:2)), useNA = 'always'))