You're loading data into a Snowflake table using 'COPY INTO'. You notice that some rows are being rejected due to data validation errors (e.g., data type mismatch, uniqueness constraint violations). You want to implement a strategy to capture these rejected rows for further analysis and correction. Which of the following approaches offers the MOST efficient and reliable method for capturing and storing the rejected rows, minimizing performance impact during the data loading process? Assume no staging table exists before loading data to production table.

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

Explanation: Only visible for ValidVCE members. You can sign-up / login (it's free).

You are developing a data transformation pipeline in Snowpark Python to aggregate website traffic data'. The raw data is stored in a Snowflake table named 'website_events' , which includes columns like 'event_timestamp' , 'user_id', 'page_urr , and 'event_type'. Your goal is to calculate the number of unique users visiting each page daily and store the aggregated results in a new table named Considering performance and resource efficiency, select all the statements that are correct:

• A. Applying a filter early in the pipeline to remove irrelevant 'event_type' values can significantly reduce the amount of data processed in subsequent aggregation steps.
• B. Defining the schema for the table before writing the aggregated results is crucial for ensuring data type consistency and optimal storage.
• C. Using followed by is an efficient approach to calculate unique users per page per day.
• D. Using is the most efficient method for writing the aggregated results to Snowflake, regardless of data size.
• E. Caching the 'website_eventS DataFrame using 'cache()' before performing the aggregation is always beneficial, especially if the data volume is large.

Explanation: Only visible for ValidVCE members. You can sign-up / login (it's free).

A data provider wants to share a large dataset (several TB) with multiple consumers. The dataset is updated daily. The provider wants to minimize the cost associated with data sharing and ensure that consumers receive consistent data'. Which of the following strategies would be the MOST cost-effective and maintainable?

• A. Create a data share and grant access to all consumers directly on the base tables.
• B. Create a data share and create a separate database for each consumer, cloning the data into each consumer's database daily.
• C. Create a data share containing external tables pointing to data stored in cloud storage (e.g., AWS S3), updated daily using a pipeline.
• D. Create a data share with views that point to the base tables. Use time travel to allow consumers to query data from a specific point in time before the daily update.
• E. Create a data share with views that point to the base tables, and clone the base tables daily into a separate 'staging' database before sharing.

Explanation: Only visible for ValidVCE members. You can sign-up / login (it's free).

You have implemented external tokenization for a sensitive data column in Snowflake using a UDF that calls an external API. After some time, you discover that the external tokenization service is experiencing intermittent outages, causing queries using the tokenized column to fail. What is the BEST approach to mitigate this issue and maintain data availability while minimizing the risk of exposing the raw data?

• A. Modify the tokenization UDF to cache tokenization mappings locally within the Snowflake environment. When the external service is unavailable, the UDF can use the cached values.
• B. Replicate the tokenized table to another Snowflake region and switch to the replica during outages of the primary region. The tokenization service is guaranteed to be available in at least one region.
• C. Implement a try-catch block within the UDF. In the catch block, return a pre-defined static token value (same value always) instead of attempting to call the external tokenization service. You can't return the raw value.
• D. Implement a masking policy on the column that returns the raw data when the tokenization UDF is unavailable, detected by catching exceptions within the policy logic.
• E. Implement a try-catch block within the UDF. In the catch block, return a pre-defined, non-sensitive default value instead of attempting to call the external tokenization service. You can't return the raw value.

Explanation: Only visible for ValidVCE members. You can sign-up / login (it's free).

You are tasked with implementing a data governance strategy in Snowflake for a large data warehouse. Your objective is to classify sensitive data columns, such as customer phone numbers and email addresses, using tags. You want to define a flexible tagging system that allows different levels of sensitivity (e.g., 'Confidential', 'Restricted') to be applied to various columns. Furthermore, you need to ensure that any data replicated to different regions maintains these classifications. Which of the following statements accurately describe best practices for implementing and maintaining data classification using tags in Snowflake, especially in a multi-region setup? Choose TWO.

• A. Tags and tag values must be uniquely defined across all schemas to avoid conflicts and ensure accurate data classification; Snowflake enforces uniqueness implicitly.
• B. Create a scheduled task that automatically identifies sensitive data based on regular expressions and applies the appropriate tags. This automates the classification process.
• C. Define tag schemas at the account level and replicate them to all regions. This ensures consistency of tag definitions across the entire organization.
• D. Always grant the ACCOUNTADMIN role to users who need to apply tags. This simplifies the process and ensures they have all necessary privileges.
• E. When replicating data between regions, the tags are automatically replicated along with the data, provided that replication is configured using database replication or failover groups including the tagging schema.

Explanation: Only visible for ValidVCE members. You can sign-up / login (it's free).