User-defined functions with Snowpark in Snowflake - Time & Space Complexity

Time Complexity: User-defined functions with Snowpark

O(n)

Understanding Time Complexity

When using user-defined functions (UDFs) with Snowpark, it is important to understand how the execution time changes as the amount of data grows.

We want to know how the number of times the UDF runs affects the total work done.

Scenario Under Consideration

Analyze the time complexity of applying a UDF to a Snowflake table column.

from snowflake.snowpark import Session
from snowflake.snowpark.functions import udf

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

@udf
def add_one(x: int) -> int:
    return x + 1

df = session.table("numbers")
df = df.select(add_one(df["value"]).alias("value_plus_one"))
df.collect()
    

This code defines a simple UDF that adds one to a number, applies it to each row in the "numbers" table, and collects the results.

Identify Repeating Operations

Look at what happens repeatedly when this code runs.

• Primary operation: The UDF is called once for each row in the table.
• How many times: Equal to the number of rows in the "numbers" table.

How Execution Grows With Input

As the number of rows increases, the UDF runs more times, directly matching the row count.

Input Size (n)	Approx. Api Calls/Operations
10	10 UDF calls
100	100 UDF calls
1000	1000 UDF calls

Pattern observation: The number of UDF calls grows directly with the number of rows.

Final Time Complexity

Time Complexity: O(n)

This means the total work grows in a straight line with the number of rows processed.

Common Mistake

[X] Wrong: "The UDF runs only once regardless of data size."

[OK] Correct: The UDF is applied to each row separately, so it runs as many times as there are rows.

Interview Connect

Understanding how UDFs scale with data size shows you can predict performance and design efficient data processing tasks.

Self-Check

"What if the UDF was applied only to a filtered subset of rows? How would the time complexity change?"