Let’s see How to Count Distinct Values of a Pandas Dataframe Column.
Consider a tabular structure as given below which has to be created as Dataframe. The columns are height, weight, and age. The records of 8 students form the rows.
| height | weight | age | |
|---|---|---|---|
| Steve | 165 | 63.5 | 20 |
| Ria | 165 | 64 | 22 |
| Nivi | 164 | 63.5 | 22 |
| Jane | 158 | 54 | 21 |
| Kate | 167 | 63.5 | 23 |
| Lucy | 160 | 62 | 22 |
| Ram | 158 | 64 | 20 |
| Niki | 165 | 64 | 21 |
The first step is to create the Dataframe for the above tabulation. Look at the code snippet below:
Python3
# import libraryimport pandas as pd# create a Dataframedf = pd.DataFrame({ 'height' : [165, 165, 164, 158, 167, 160, 158, 165], 'weight' : [63.5, 64, 63.5, 54, 63.5, 62, 64, 64], 'age' : [20, 22, 22, 21, 23, 22, 20, 21]}, index = ['Steve', 'Ria', 'Nivi', 'Jane', 'Kate', 'Lucy', 'Ram', 'Niki'])# show the Dataframeprint(df) |
Output:
height weight age
Steve 165 63.5 20
Ria 165 64.0 22
Nivi 164 63.5 22
Jane 158 54.0 21
Kate 167 63.5 23
Lucy 160 62.0 22
Ram 158 64.0 20
Niki 165 64.0 21
Using for loop
The Dataframe has been created and one can hard coded using for loop and count the number of unique values in a specific column. For example In the above table, if one wishes to count the number of unique values in the column height. The idea is to use a variable cnt for storing the count and a list visited that has the previously visited values. Then for loop that iterates through the ‘height’ column and for each value, it checks whether the same value has already been visited in the visited list. If the value was not visited previously, then the count is incremented by 1.
Below is the implementation:
Python3
# import libraryimport pandas as pd# create a Dataframedf = pd.DataFrame({ 'height' : [165, 165, 164, 158, 167, 160, 158, 165], 'weight' : [63.5, 64, 63.5, 54, 63.5, 62, 64, 64], 'age' : [20, 22, 22, 21, 23, 22, 20, 21]}, index = ['Steve', 'Ria', 'Nivi', 'Jane', 'Kate', 'Lucy', 'Ram', 'Niki'])# variable to hold the countcnt = 0# list to hold visited valuesvisited = []# loop for counting the unique# values in heightfor i in range(0, len(df['height'])): if df['height'][i] not in visited: visited.append(df['height'][i]) cnt += 1print("No.of.unique values :", cnt)print("unique values :", visited) |
Output :
No.of.unique values : 5
unique values : [165, 164, 158, 167, 160]
But this method is not so efficient when the Dataframe grows in size and contains thousands of rows and columns. To give an efficient there are three methods available which are listed below:
- pandas.unique()
- Dataframe.nunique()
- Series.value_counts()
Method 1: Using unique()
The unique method takes a 1-D array or Series as an input and returns a list of unique items in it. The return value is a NumPy array and the contents in it based on the input passed. If indices are supplied as input, then the return value will also be the indices of the unique value.
Syntax: pandas.unique(Series)
Example:
Python3
# import libraryimport pandas as pd# create a Dataframedf = pd.DataFrame({ 'height' : [165, 165, 164, 158, 167, 160, 158, 165], 'weight' : [63.5, 64, 63.5, 54, 63.5, 62, 64, 64], 'age' : [20, 22, 22, 21, 23, 22, 20, 21]}, index = ['Steve', 'Ria', 'Nivi', 'Jane', 'Kate', 'Lucy', 'Ram', 'Niki'])# counting unique valuesn = len(pd.unique(df['height']))print("No.of.unique values :", n) |
Output:
No.of.unique values : 5
Method 2: Using Dataframe.nunique()
This method returns the count of unique values in the specified axis. The syntax is :
Syntax: Dataframe.nunique (axis=0/1, dropna=True/False)
Example:
Python3
# import libraryimport pandas as pd# create a Dataframedf = pd.DataFrame({ 'height' : [165, 165, 164, 158, 167, 160, 158, 165], 'weight' : [63.5, 64, 63.5, 54, 63.5, 62, 64, 64], 'age' : [20, 22, 22, 21, 23, 22, 20, 21]}, index = ['Steve', 'Ria', 'Nivi', 'Jane', 'Kate', 'Lucy', 'Ram', 'Niki'])# check the values of# each row for each columnn = df.nunique(axis=0)print("No.of.unique values in each column :\n", n) |
Output:
No.of.unique values in each column :
height 5
weight 4
age 4
dtype: int64
To get the number of unique values in a specified column:
Syntax: Dataframe.col_name.nunique()
Example:
Python3
# import libraryimport pandas as pd# create a Dataframedf = pd.DataFrame({ 'height' : [165, 165, 164, 158, 167, 160, 158, 165], 'weight' : [63.5, 64, 63.5, 54, 63.5, 62, 64, 64], 'age' : [20, 22, 22, 21, 23, 22, 20, 21]}, index = ['Steve', 'Ria', 'Nivi', 'Jane', 'Kate', 'Lucy', 'Ram', 'Niki'])# count no. of unique# values in height columnn = df.height.nunique()print("No.of.unique values in height column :", n) |
Output:
No.of.unique values in height column : 5
Method 3: Using Series.value_counts()
This method returns the count of all unique values in the specified column.
Syntax: Series.value_counts(normalize=False, sort=True, ascending=False, bins=None, dropna=True)
Example:
Python3
# import libraryimport pandas as pd# create a Dataframedf = pd.DataFrame({ 'height' : [165, 165, 164, 158, 167, 160, 158, 165], 'weight' : [63.5, 64, 63.5, 54, 63.5, 62, 64, 64], 'age' : [20, 22, 22, 21, 23, 22, 20, 21]}, index = ['Steve', 'Ria', 'Nivi', 'Jane', 'Kate', 'Lucy', 'Ram', 'Niki'])# getting the list of unique valuesli = list(df.height.value_counts())# print the unique value countsprint("No.of.unique values :", len(li)) |
Output:
No.of.unique values : 5
