Get Unique Values from a Column in Pandas DataFrame in Python

Introduction

One of the most powerful data manipulation libraries in Python is Pandas. In addition, it provides a range of structured data functions. Actually about the DataFrames in particular, one often just needs to consider only unique values for a certain column. In this chapter we examine some of the methods for obtaining all those elements you require.

Understanding Pandas DataFrame

So first off, we'll skip ahead a bit and quickly cover some basic facts about Pandas DataFrames. That is, before getting into the technical details of how to get unique values. A DataFrame is a two-dimensional labeled data table, with rows and columns. It was custom made for data work, sitting on the shoulders of NumPy.

Output:

Name  Age           City
0    Alice   25       New York
1      Bob   30  San Francisco
2    Alice   25       New York
3  Charlie   35    Los Angeles
4      Bob   30  San Francisco

Method 1: Using 'unique()' Method

Pandas unique() method is an efficient way to get the unique elements of a column. It returns an array containing only unique values, in the order that they appear in DataFrame.

Output:

Unique Names: ['Alice' 'Bob' 'Charlie']

In this piece of code, 'unique()' in Pandas gets the unique values from df['Name'] column. unique_names is an array of the original names displayed in order. This print statement displays these special names.

Method 2: Using 'value_counts()' Method

In addition to giving unique values, the 'value_counts()' method also counts their occurrences. If you want to know how many times each unique element occurs in a given column, it can be very useful.

Output:

Name Counts:
 Bob        2
Alice      2
Charlie    1
Name: Name, dtype: int64

Here, the 'value_counts()' method is used to extract both unique names and their counts from that Name column. The result of name_counts is a Pandas Series providing the frequency distribution for all unique names.

Method 3: Using 'drop_duplicates()' Method

A second way to get unique values is the 'drop_duplicates()' method. Unlike unique(), this method returns a new DataFrame containing no duplicates.

Output:

DataFrame with Unique Names:
     Name  Age           City
0  Alice   25       New York
1    Bob   30  San Francisco
3 Charlie   35    Los Angeles

Drop duplicate rows based on the 'Name' column (unique_df) using drop_duplicates(). As a result, our DataFrame retains only the first instance of each unique name and we have set which is clean.

Method 4: Applying a Set

By definition, Python's set stores only unique elements. If we change a column to set, then finding all the distinct values is easy.

Output:

Unique Cities: {'San Francisco', 'Los Angeles', 'New York'}

This tiny piece of code turns the 'City' column into a set (unique_cities). Since sets, by definition, contain only non-repeated elements this procedure finds the city names that are different from DataFrame and prints them.

Method 5: Using 'nunique()' Method

The method 'nunique()' returns the number of unique elements in a column. It's especially good when what you want is a count of unique values but without having to enumerate them.

Output:

Number of Unique Names: 3

'nunique()' calculates the number of unique names in that column, returning a single numeric value (num_unique_names). The print statement shows the number of unique names.

Method 6: Custom Functions for Unique Values

In other cases, you will have to introduce custom logic of your own in order determine unique values. It could also involve using a function which checks for uniqueness based on certain criteria.

Output:

Unique Names based on Custom Logic: []

A custom function ('custom_unique_check') is defined to check uniqueness according to a specific standard, for example that the name be even in length. This function is then applied to 'Name' using the 'apply() method, and the resulting DataFrame contains all values meeting our custom condition. The names that meet the criterion are then printed in a unique form.

Conclusion

In this exhaustive guide we went over how to extract novel values from a column in Pandas DataFrame. Whether your precision requirements dictate the use of built-in methods such as 'unique()', 'value_counts()' and/or, drop duplicates (), or you choose to write custom functions, Pandas offers a range options for meeting all manner of needs.

Knowing these skills are essential for data cleansing, preprocessing and analysis work which enable you to appreciate what makes your datasets special. As you continue your work with Pandas DataFrames, learning these methods will make it easier to break down and extract information from your data.