User-defined functions with Snowpark in Snowflake - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is a user-defined function (UDF) in Snowpark?

A user-defined function (UDF) in Snowpark is a custom function that you write to perform specific operations on data inside Snowflake. It lets you add your own logic to process data beyond built-in functions.

beginner

How do you create a simple scalar UDF using Snowpark in Python?

You define a Python function with the logic you want, then register it as a UDF using Snowpark's session.udf.register method. This makes the function available to use in SQL queries.

intermediate

What types of UDFs can you create with Snowpark?

You can create scalar UDFs that return a single value, and table UDFs that return a table (multiple rows and columns). Snowpark supports both types to handle different data processing needs.

intermediate

Why use Snowpark UDFs instead of SQL alone?

Snowpark UDFs let you write complex logic in familiar programming languages like Python or Java, which can be easier and more powerful than writing complex SQL. They run close to the data for better performance.

beginner

What is the role of the Snowflake session object when working with Snowpark UDFs?

The session object manages the connection to Snowflake and provides methods to register UDFs, run queries, and manage data. It is the main entry point for working with Snowpark features.

What does a scalar UDF return in Snowpark?

A A single value

B Multiple rows and columns

C A SQL query

D A database connection

Scalar UDFs return a single value for each input row.

Which programming language can you use to write Snowpark UDFs?

A CSS

B HTML

C Markdown

D Python

Snowpark supports writing UDFs in Python (and also Java and Scala).

How do you make a Python function available as a UDF in Snowflake using Snowpark?

A Save it as a CSV file

B Write it in SQL

C Register it with session.udf.register

D Use a web browser

You register the Python function with session.udf.register to create a UDF.

What is a table UDF in Snowpark?

A A function that returns multiple rows and columns

B A function that returns a single value

C A function that connects to the internet

D A function that deletes data

Table UDFs return a table with multiple rows and columns.

Why is it beneficial to run UDFs close to the data in Snowflake?

A It requires more storage

B It improves performance by reducing data movement

C It slows down queries

D It makes the data disappear

Running UDFs close to the data reduces data transfer and speeds up processing.

Explain how to create and use a simple scalar UDF with Snowpark in Python.

Think about writing a function and making it available in Snowflake.

Describe the difference between scalar and table UDFs in Snowpark.

Consider what kind of output each UDF produces.