How to get the table counts for unique values in column

Question

I have this matrix mymat. I know I can do table((mymat[,"col1"]) to get the number of each item in col1. However, I only want the count if there are unique values in col2. For mymat below, I want this result:

app  gg  chh
 1    2   1

mymat

col1   col2
app    d
app    d
gg     e
gg     f 
gg     e
chh    f
chh    f
chh    f

Answer 1

You can use unique to subset the data (works for matrix and data.frame) and then call table:

table(unique(mymat)[,1])

This returns

# app chh  gg 
#   1   1   2 

Answer 2

You can use duplicated to subset the data and then call table:

table(subset(df, !duplicated(paste(col1, col2)), select = col1))
#app chh  gg 
#  1   1   2 

As a second option, here's a dplyr approach:

library(dplyr)
distinct(df) %>% count(col1)   # or distinct(df, col1, col2) if you have other columns
#Source: local data frame [3 x 2]
#
#    col1     n
#  (fctr) (int)
#1    app     1
#2    chh     1
#3     gg     2

Answer 3

This is counting the nonzeros in table()-result

rowSums(table(df$col1, df$col2)!=0)

result:

app chh  gg 
  1   1   2 

data used:

df <- read.table(header=TRUE, text=
"col1   col2
app    d
app    d
gg     e
gg     f 
gg     e
chh    f
chh    f
chh    f")