How to Count Unique Values by Group in R (With Examples)

---

You can use the following methods to count the number of unique values by group in R:

Method 1: Using Base R

results <- aggregate(data=df, values_var~group_var, function(x) length(unique(x)))

Method 2: Using dplyr

library(dplyr)

results <- df %>%
  group_by(group_var) %>%
  summarize(count = n_distinct(values_var))

Method 3: Using data.table

library(data.table)

df <- data.table(df)
results <- df[ , .(count = length(unique(values_var))), by = group_var]

Each method returns the exact same result, but the base R method tends to be significantly slower when working with large data frames.

The following examples show how to use each of these methods in practice with the following data frame:

#create data frame
df <- data.frame(team=c('A', 'A', 'A', 'A', 'B', 'B', 'C', 'C', 'C'),
                 points=c(10, 10, 14, 14, 18, 19, 20, 20, 20))

#view data frame
df

  team points
1    A     10
2    A     10
3    A     14
4    A     14
5    B     18
6    B     19
7    C     20
8    C     20
9    C     20

Method 1: Count Unique Values by Group Using Base R

The following code shows how to count the number of distinct points values for each team using base R:

#count unique points values by team
results <- aggregate(data=df, points~team, function(x) length(unique(x)))

#view results
results

  team points
1    A      2
2    B      2
3    C      1

From the output we can see:

• There are 2 unique points values for team A.
• There are 2 unique points values for team B.
• There is 1 unique points value for team C.

Method 2: Count Unique Values by Group Using dplyr

The following code shows how to count the number of distinct points values for each team using dplyr:

library(dplyr)

#count unique points values by team
results <- df %>%
  group_by(team) %>%
  summarize(count = n_distinct(points))

#view results
results

# A tibble: 3 x 2
  team  count
1 A         2
2 B         2
3 C         1

Notice that these results match the ones from the base R method.

Method 3: Count Unique Values by Group Using data.table

The following code shows how to count the number of distinct points values for each team using data.table:

library(data.table)

#convert data frame to data table
df <- data.table(df)

#count unique points values by team 
results <- df[ , .(count = length(unique(points))), by = team]

#view results
results

   team count
1:    A     2
2:    B     2
3:    C     1

Notice that these results match the ones from the previous two methods.

The following tutorials explain how to perform other common operations using dplyr: