User-defined functions with Snowpark in Snowflake - Mini Project: Build & Apply

User-defined functions with Snowpark

📖 Scenario: You are working with Snowflake's Snowpark to create a user-defined function (UDF) that processes data inside the database. This helps you run custom logic close to your data for faster and scalable processing.

🎯 Goal: Build a simple user-defined function (UDF) using Snowpark that takes a string input and returns the string in uppercase.

📋 What You'll Learn

Create a Snowpark session variable called session.

Define a Python function called to_upper that takes one string argument.

Register the to_upper function as a UDF named uppercase_udf in Snowflake.

Use the UDF in a Snowpark DataFrame to transform a column of strings to uppercase.

💡 Why This Matters

🌍 Real World

User-defined functions let you run custom code close to your data in Snowflake, improving performance and enabling complex logic inside the database.

💼 Career

Knowing how to create and use UDFs with Snowpark is valuable for data engineers and analysts who want to extend Snowflake's capabilities with custom processing.

1

Create a Snowpark session

Create a Snowpark session variable called session using Session.builder.configs() with an empty dictionary {} and then call .create().

Snowflake

# Create a Snowpark session variable called session
# Your code here

Need a hint?

Use Session.builder.configs({}).create() to create the session.

2

Define the Python function to convert strings to uppercase

Define a Python function called to_upper that takes one argument input_str and returns input_str.upper().

Snowflake

from snowflake.snowpark import Session

session = Session.builder.configs({}).create()

# Define the function to_upper that returns the uppercase of input_str
# Your code here

Need a hint?

Use the upper() method on the string argument.

3

Register the Python function as a Snowflake UDF

Register the function to_upper as a Snowflake UDF named uppercase_udf using session.udf.register() with to_upper and name='uppercase_udf'.

Snowflake

from snowflake.snowpark import Session

session = Session.builder.configs({}).create()

def to_upper(input_str):
    return input_str.upper()

# Register to_upper as a UDF named uppercase_udf
# Your code here

Need a hint?

Use session.udf.register() with the function and the name.

4

Use the UDF in a Snowpark DataFrame

Create a Snowpark DataFrame called df from a list of strings [('hello',), ('world',)] with column name 'word'. Then create a new DataFrame called df_upper by selecting the column 'word' and applying the UDF uppercase_udf on it as uppercase_word.

Snowflake

from snowflake.snowpark import Session

session = Session.builder.configs({}).create()

def to_upper(input_str):
    return input_str.upper()

uppercase_udf = session.udf.register(to_upper, name='uppercase_udf')

# Create DataFrame df and apply uppercase_udf to create df_upper
# Your code here

Need a hint?

Use session.create_dataframe() with the list and schema, then use select() with the UDF and alias().