You have a SQL query stored in a file named 'query.sqr which contains several complex analytical calculations. The query depends on a Snowpark 'session' object already established. You want to create a Snowpark DataFrame from the result of this query. Which of the following code snippets achieves this with optimal performance and readability, assuming correct file access permissions?

• A.
  
• B.
  
• C.
  
• D.
  
• E.
  

You have a Snowpark DataFrame containing customer data'. You need to create a stored procedure that accepts the DataFrame and a list of column names as input and returns a new DataFrame containing only the specified columns. Which of the following approaches correctly implement this functionality and handles data types effectively (Select all that apply)?

• A.
  
• B.
  
• C.
  
• D.
  
• E.
  

Consider the following Snowpark Python code snippet:

• A. This code requires a configured Anaconda environment to run successfully.
• B. The function will retrieve all rows from the 'customers' table and store them in a local Pandas DataFrame before applying the function.
• C. The code demonstrates the Snowpark architecture, where transformations are translated into SQL and executed in Snowflake's engine. Only the final 'collect()' brings the results back to the client.
• D.
  
• E. The 'upper()' function will be executed on the client-side (where the Python code is running) for each row in the 'customers' table.

You have a Snowpark application that reads data from a large Snowflake table and performs several transformations. During testing, you observe that the application's performance is inconsistent, with some runs taking significantly longer than others, even with the same input data'. You suspect that data locality might be a contributing factor. What steps can you take within your Snowpark application to investigate and potentially improve data locality and performance consistency?

• A. Ensure the Snowpark session is configured with a large enough warehouse size to minimize data spilling to disk.
• B. Disable Snowflake's result cache. This ensures that the application always reads the most recent data from disk, regardless of performance impact.
• C. Implement caching using , combined with a targeted 'repartition()' to ensure that frequently accessed data is readily available in memory close to the processing nodes.
• D. Use to redistribute the data across the cluster based on a relevant key. This can improve data locality for subsequent operations.
• E. Enable Snowflake's automatic clustering on the underlying table if it's not already enabled. This will physically organize the data on disk based on the clustering key.

You have two Snowpark DataFrames: containing customer information and 'orders_df containing order details. You need to merge these DataFrames based on the column to create a unified view. The 'customers_df may contain duplicate values. The contains recent orders. You want to use 'merge' with an 'UPDATE action based on 'customer id' and a 'WHEN NOT MATCHED BY TARGET action to insert new customer records from 'customers df into the 'orders df table that do not exist.

• A. The 'merge' operation will succeed, processing each matching record sequentially. If more than one row matches it will process based on order in the dataframe
• B. The 'merge' operation will fail immediately upon encountering the first duplicate key during the update operation. Consider stage the source dataframe and use an appropriate dedup option with merge.
• C. The merge operation cannot handle duplicate keys in the source DataFrame without pre-aggregation.
• D. The 'UPDATE action in 'merge' will update all matching rows in the target table based on the join condition, regardless of duplicates in the source.
• E. You must use on the source DataFrame ('customers_df) before using 'merge' to ensure only one matching row exists per customer.