SnowPro Associate: Platform Question 34
Single answer○ PythonA data engineering team wants to load a pandas DataFrame from a Python application into a Snowflake table with the least amount of custom SQL and the best support for handling larger batches efficiently. The application already uses the Snowflake Connector for Python. Which approach should the team use?
- A
Use the connector's write_pandas() helper to upload the DataFrame and load it into the target table.
- B
Convert the DataFrame to a list of tuples and issue one INSERT statement per row using cursor.execute().
- C
Save the DataFrame locally as a CSV file and use pandas.read_sql() to load it into Snowflake.
- D
Use the Python worksheet runtime to automatically persist the DataFrame into a table without specifying a target table.
Show answer and explanation
Correct answer: A
Explanation
For Python-based ingestion of pandas DataFrames into Snowflake, a practical best practice is to use the Snowflake Connector for Python helper write_pandas(). This aligns with Snowflake documentation for pandas-oriented workflows and avoids inefficient row-by-row INSERT patterns. The key applied concept is choosing the right Python integration point for bulk DataFrame loading: use connector-supported DataFrame loading utilities rather than low-throughput manual inserts or APIs intended for reading data. Candidates should recognize that Python in Snowflake-related workflows often involves the Snowflake Connector for Python, Snowpark APIs, and worksheet environments, but each has specific purposes and does not eliminate the need to explicitly write data to a target table.
- A. Correct.
Correct. The Snowflake Connector for Python provides write_pandas() specifically to load pandas DataFrames into Snowflake tables efficiently. It is designed to reduce custom SQL and is the recommended high-level helper when working with pandas in Python applications that already use the connector. It is generally more practical and scalable than row-by-row inserts for larger batches.
- B. Incorrect.
Incorrect. While inserting rows individually with cursor.execute() can work functionally, it is inefficient for larger DataFrames and requires more custom logic. This approach increases network round trips and is not the best option when the goal is minimal custom SQL and efficient batch loading.
- C. Incorrect.
Incorrect. pandas.read_sql() is used to read query results from a database into a DataFrame, not to load a local CSV or DataFrame into Snowflake. This option confuses data extraction from Snowflake with data loading into Snowflake.
- D. Incorrect.
Incorrect. Python worksheets are a Snowflake development interface, but they do not automatically persist a pandas DataFrame into a table without an explicit load/write operation and target object. A table destination must still be specified through the appropriate API or SQL process.