DEA-C02 Practice Test 2 — 30 Questions

Question 1

A user runs a query twice in succession with no data changes. The second query completes in near-zero time. What feature is responsible for this performance?

Accepted Answer

Query Result Cache.. The Query Result Cache stores the output of every query for a rolling 24-hour period. When a subsequent query matches the exact SQL text and the underlying data in the source tables has not changed, Snowflake returns the result from the cache. This mechanism avoids the need to spin up compute resources or scan data, providing an instantaneous response for repeated queries, which is a major efficiency feature for frequently accessed reports.

Answer

Warehouse Local Cache.

Answer

Metadata Cache.

Answer

Search Optimization Service.

Question 2

What is the primary benefit of using a file format object in Snowflake when dealing with multiple stages and load jobs?

Accepted Answer

It centralizes file configuration to ensure consistency across multiple load jobs.. File format objects provide a centralized, reusable definition of file settings (like delimiter, compression, or error handling). By referencing a single object in multiple COPY commands, an engineer ensures consistency across the entire data pipeline. This eliminates drift and makes it easy to update the configuration for all load jobs at once, reducing maintenance and preventing errors associated with manually defining parameters repeatedly across many different SQL statements.

Answer

It automatically compresses data before it is uploaded to the cloud.

Answer

It encrypts the data at rest in the stage before loading.

Answer

It stores the credentials for the external stages for easier access.

Question 3

A data engineer is tuning a query that joins two large tables. The query is performing a 'Remote Disk Spilling' operation. Which optimization strategy is most effective to resolve this?

Accepted Answer

Scale up the warehouse to a larger size.. Remote disk spilling occurs when the memory allocated to the virtual warehouse is insufficient to hold the intermediate result sets of a join or aggregation. By increasing the size of the warehouse, the memory available per node doubles, allowing larger datasets to be processed entirely in-memory. This significantly reduces latency associated with I/O operations and speeds up complex join operations that frequently cause spilling in standard configurations.

Answer

Increase the multi-cluster warehouse scaling policy to 'Maximized'.

Answer

Enable Result Cache by setting USE_CACHED_RESULT to TRUE.

Answer

Implement a materialized view on the joined columns.

Question 4

Which approach is most efficient for transforming a large volume of data in Snowflake when the logic requires complex window functions and stateful processing?

Accepted Answer

Using set-based SQL transformations in a materialized view or dynamic table.. For large-scale, complex transformations, using SQL within a View, Dynamic Table, or a CTAS (Create Table As Select) operation is highly efficient because it runs directly on the Snowflake compute engine. Snowflake's query optimizer is highly tuned for window functions and complex joins, allowing it to distribute these operations across all nodes in the warehouse, providing superior performance compared to row-by-row procedural processing.

Answer

Using a Python script to fetch data, process locally, and upload.

Answer

Using a single Stored Procedure with row-by-row cursor processing.

Answer

Using a User-Defined Function (UDF) for every column transformation.

Question 5

An organization is unloading data from a Snowflake table to an external S3 bucket for use in a machine learning pipeline. Which TWO practices will optimize the performance and manageability of the exported files? (Select TWO)

Accepted Answer

Use the PARTITION BY clause in the COPY INTO <location> statement to organize files into subfolders.. Optimizing data unloading requires balancing file size and organizational structure. Using the PARTITION BY clause ensures that the exported files are organized into a logical folder hierarchy, which facilitates faster data discovery for downstream tools. Additionally, controlling the MAX_FILE_SIZE ensures that the files are not too large for the consuming application to process efficiently in memory while maintaining parallelism.

Answer

Set the SINGLE parameter to TRUE to ensure all data is consolidated into one large file.

Answer

Disable compression using the COMPRESSION = NONE parameter to reduce the CPU load on the warehouse.

Answer

Always use the OVERWRITE = TRUE parameter to prevent the command from failing if files already exist.

Question 6

Which approach is most effective for managing governance policies across a large, multi-schema data warehouse environment?

Accepted Answer

Deploy policies to a centralized database and schema.. Centralizing governance objects in a dedicated 'GOVERNANCE' database is the best practice. By keeping policies, tags, and data classification results in a single, well-controlled schema, organizations ensure consistency and ease of maintenance. This centralized approach simplifies access control for the security team and allows for easier auditing of policy changes compared to scattering governance artifacts across disparate, business-specific databases or schemas.

Answer

Create policies within each individual schema.

Answer

Use a single shared role for all policy management.

Answer

Manually recreate all policies in every environment.

Question 7

Which Snowflake feature should be used to move data from a private cloud network to Snowflake without traversing the public internet?

Accepted Answer

AWS PrivateLink or Azure Private Link.. For enterprises with strict security requirements, using AWS PrivateLink or Azure Private Link allows for a secure, private connection between the cloud network and Snowflake. This is vital for data movement in highly regulated industries. By eliminating the need for public internet exposure, organizations minimize their attack surface and comply with networking policies, ensuring that sensitive data remains within a controlled, private infrastructure environment throughout the entire transit process.

Answer

SnowSQL with VPN tunneling.

Answer

Snowflake Data Exchange.

Answer

Snowpipe with an external stage.

Question 8

What is the primary purpose of the 'VALIDATION_MODE' parameter in the COPY INTO <table_name> statement?

Accepted Answer

To check the files for errors and return the results without loading the data into the table.. The VALIDATION_MODE parameter allows data engineers to test their COPY commands without actually committing any data to the table. It parses the files and identifies errors, which is a critical step in data movement pipelines to ensure that file formats and mappings are correct before performing a potentially expensive or large-scale ingestion.

Answer

To automatically correct data type mismatches during the load process.

Answer

To verify that the person executing the command has the correct IAM permissions on the S3 bucket.

Answer

To compare the data in the stage with the data already in the table to find duplicates.

Question 9

Which TWO of the following are true regarding the use of Snowflake Data Classification?

Accepted Answer

The process uses system-defined tags to label sensitive data.. Snowflake Data Classification automatically identifies PII and sensitive data within tables, assigning system tags to columns. This process significantly reduces the manual effort required for governance. It works by scanning sample data and applying semantic labels, which can then be used to trigger automated protection measures like masking policies. Understanding how this automates compliance is essential for any modern data engineer managing large, dynamic datasets.

Answer

Data classification can only be applied to existing tables.

Answer

Classification results are stored in the user's local file system.

Answer

Classification requires the data to be in a flat file format.

Question 10

A data engineer is reviewing a query profile for a query that performs a large join. The profile shows a high percentage of time spent in the 'Join' node with significant 'Bytes spilled to local storage'. The engineer wants to reduce local spilling without changing the query. Which action is most appropriate?

Accepted Answer

Increase the warehouse size to provide more memory for the join operation.. Local spilling during a join indicates that the join operation is exceeding the memory available on a warehouse node. Increasing the warehouse size provides more memory per node, which can accommodate larger intermediate results and reduce spilling. Filtering, clustering, or changing join types may have indirect benefits but do not directly resolve the memory pressure. Therefore, scaling up the warehouse is the most appropriate action.

Answer

Reduce the size of the smaller table in the join by filtering it.

Answer

Add a clustering key on the join column of the larger table.

Answer

Change the join type from a hash join to a sort-merge join.

Question 11

A data engineer needs to unload a large table to an external stage in Parquet format. The table contains a column with sensitive data that must be masked in the unloaded files. The engineer wants to use a secure view that applies a masking policy, and then unload from that view. Which statement accurately describes the behavior when unloading from a secure view?

Accepted Answer

Unloading from a secure view is allowed, and the masking policies defined in the view are applied to the unloaded data.. COPY INTO <location> supports unloading from secure views, and the secure view's definition, including masking and row access policies, is enforced during the unload. This means sensitive columns can be masked in the output files. The other options either deny support, claim masking is bypassed, or impose unnecessary role requirements. The correct behavior is that policies are applied, ensuring data security.

Answer

Unloading from a secure view is allowed, but the masking policies are not applied; the underlying data is unloaded unmasked.

Answer

Unloading from a secure view is not supported; COPY INTO <location> requires a base table or a regular view.

Answer

Unloading from a secure view requires the ACCOUNTADMIN role and the masking policies are ignored.

Question 12

A data engineer needs to protect a set of permanent tables in a Snowflake Enterprise Edition account. The requirement is to allow querying of historical versions for up to 30 days and to allow recovery of dropped tables within that same window. Which two actions should the engineer take to meet these requirements? (Choose two.)

Accepted Answer

Confirm the account is on Enterprise Edition, which supports retention windows longer than the Standard Edition one-day maximum.. Meeting a 30-day queryable-history requirement on permanent tables requires Enterprise Edition, which raises the maximum retention from one day to ninety days, and an explicit DATA_RETENTION_TIME_IN_DAYS setting of 30 on each table. Transient tables cap retention at one day, Fail-safe is non-queryable and fixed at seven days, and there is no configurable FAILSAFE_DAYS parameter.

Answer

Convert the permanent tables to transient tables to increase the maximum allowed retention window.

Answer

Rely on Fail-safe to provide the 30-day queryable history window for dropped tables.

Answer

Enable Fail-safe on the tables by setting FAILSAFE_DAYS to 30 using ALTER TABLE.

Question 13

A data engineer manages a transient table named STG_EVENTS in a Snowflake Enterprise Edition account. The table has a DATA_RETENTION_TIME_IN_DAYS setting of 1. The engineer wants Time Travel to retain data for 30 days for this table. What will happen if the engineer executes ALTER TABLE STG_EVENTS SET DATA_RETENTION_TIME_IN_DAYS = 30?

Accepted Answer

The statement will fail with an error because transient tables support a maximum Time Travel retention of 1 day.. Transient tables are limited to a maximum Time Travel retention of 1 day. Attempting to set a higher value such as 30 days causes Snowflake to reject the ALTER TABLE statement with an error. To achieve 30 days of Time Travel, the table must be a permanent table in an account with Enterprise Edition or higher. The table type is the deciding factor.

Answer

The retention period will be set to 30 days, but only if the table has not been cloned previously.

Answer

The retention period will be set to 30 days but Fail-safe will still be disabled for the table.

Answer

The retention period will be set to 30 days, and Time Travel will retain data for 30 days.

Question 14

A data engineer is implementing a data classification process using Snowflake's Data Classification feature. The engineer wants to automatically classify columns containing sensitive data and then use the results to apply masking policies. After running the classification, the engineer notices that some columns that should be classified as 'EMAIL' are not being tagged. The engineer has verified that the data contains valid email addresses. What is the most likely reason for the missing classification?

Accepted Answer

The Data Classification feature only samples a subset of rows, and the sample did not include enough email addresses to meet the threshold.. Data Classification uses sampling to analyze column data. If the sample does not contain a sufficient number of email addresses, the column may not be classified as EMAIL. This is a known limitation. The other options are incorrect because classification does not depend on column names, can handle NULLs, and is not restricted to a specific schema.

Answer

The Data Classification feature requires that the column name contains the word 'EMAIL' to classify it as EMAIL.

Answer

The Data Classification feature cannot classify columns that contain NULL values.

Answer

The Data Classification feature is only available for columns in the PUBLIC schema.

Question 15

A data engineer is auditing storage consumption in a Snowflake account. A large permanent table named SALES_FACT has DATA_RETENTION_TIME_IN_DAYS set to 60. Over the past month, ETL jobs have performed several full-table UPDATE operations that rewrote most micro-partitions. The engineer notices that the table's Time Travel storage has grown significantly even though the visible row count is unchanged. What is the most likely cause of the increased Time Travel storage?

Accepted Answer

Time Travel storage grows because each UPDATE creates new micro-partitions while prior versions are retained for the 60-day retention window.. Time Travel storage reflects historical micro-partition versions retained for the configured window. Each UPDATE that rewrites micro-partitions creates new versions, and the prior versions are stored for 60 days. Repeated full-table updates therefore accumulate substantial historical data, increasing Time Travel storage even when the visible row count is stable.

Answer

Time Travel storage grows because the 60-day retention forces Snowflake to store a full compressed copy of the table in Fail-safe.

Answer

Time Travel storage grows because Snowflake copies the entire table into a separate Time Travel schema on each UPDATE.

Answer

Time Travel storage grows because the account has exceeded its storage quota, causing Snowflake to duplicate micro-partitions as a redundancy measure.

Question 16

A data engineer is designing a transformation that requires unpivoting a wide table with columns `q1_sales`, `q2_sales`, `q3_sales`, and `q4_sales` into a long format with columns `quarter` and `sales`. The table has millions of rows. Which Snowflake construct should be used to achieve this efficiently?

Accepted Answer

Use the `UNPIVOT` clause in a SELECT statement.. The UNPIVOT clause is specifically designed to transform columns into rows. It is a native Snowflake feature that is optimized for performance on large tables. Using UNION ALL or LATERAL FLATTEN can achieve similar results but with more complexity and potentially less efficiency. PIVOT does the reverse and is not appropriate here.

Answer

Use the `PIVOT` clause in a SELECT statement.

Answer

Use a `LATERAL FLATTEN` on an array constructed from the quarter columns.

Answer

Use a series of `UNION ALL` SELECT statements, one for each quarter column.

Question 17

A data engineer needs to continuously ingest JSON event files landing in an Amazon S3 bucket. The ingestion must be automatic, near real-time, and must reuse the same transformation logic already validated in a COPY INTO statement. The team wants the least administrative overhead. Which Snowflake feature should they configure?

Accepted Answer

A Snowpipe configured with COPY INTO from the S3 stage. Snowpipe is the Snowflake service built for continuous, event-driven loading. It runs COPY INTO statements automatically when new files arrive and can embed transformation logic, so the validated COPY logic is reused. Because it uses serverless compute triggered by cloud event notifications, it provides near real-time ingestion with very little operational management, fitting the stated constraints.

Answer

A scheduled task that runs COPY INTO every five minutes against the S3 stage

Answer

An external table defined over the S3 stage with automatic refresh enabled

Answer

A materialized view created over an external table pointing at the S3 stage

Question 18

A data engineer is using Snowpark for Python to transform data. The engineer needs to perform a join between two DataFrames and then apply a filter based on a column from the joined result. Which TWO methods are appropriate for this transformation? (Choose two.)

Accepted Answer

Use `DataFrame.join()` to combine the DataFrames, then `DataFrame.where()` to apply the condition.. In Snowpark, join() combines two DataFrames, and filter() or its alias where() applies a row-wise condition. These methods build a lazy query plan that executes on Snowflake. Other methods like merge(), union(), or select() do not perform a join followed by a row filter, so they are not appropriate for this transformation.

Answer

Use `DataFrame.union()` to combine the DataFrames, then `DataFrame.filter()` to apply the condition.

Answer

Use `DataFrame.join()` to combine the DataFrames, then `DataFrame.select()` to apply the condition.

Answer

Use `DataFrame.merge()` to combine the DataFrames, then `DataFrame.filter()` to apply the condition.

Question 19

A data engineer is building a transformation pipeline that must parse semi-structured log data stored in a VARIANT column named log_data. The JSON structure contains a nested array under the key 'events', and each element in the array has a 'timestamp' field. The engineer needs to produce one row per event with the timestamp extracted as a TIMESTAMP_NTZ value. Which SQL construct should be used to achieve this transformation efficiently?

Accepted Answer

Use the FLATTEN function with the LATERAL keyword to explode the 'events' array, and access the 'timestamp' field using the colon notation, casting it to TIMESTAMP_NTZ.. The FLATTEN function with LATERAL is the standard Snowflake construct for exploding nested arrays in VARIANT columns. It produces one row per array element, and the colon notation allows direct access to nested fields. Casting the extracted value to TIMESTAMP_NTZ ensures the correct data type for downstream processing. This method is efficient and leverages Snowflake's native semi-structured data handling.

Answer

Use the GET_PATH function to directly extract the 'timestamp' field from the array without exploding it, relying on implicit casting to TIMESTAMP_NTZ.

Answer

Use the OBJECT_CONSTRUCT function to rebuild the JSON, then use a JavaScript UDF to iterate over the array and return each timestamp.

Answer

Use the PARSE_JSON function to convert the VARIANT to a string, then use SPLIT_TO_TABLE to separate the array elements, and finally extract the timestamp with regular expressions.

Question 20

A data engineer needs to create a development copy of a 5 TB production table named FACT_SALES for query testing. The team wants the copy to share the same underlying micro-partitions as the source so no additional storage is consumed at creation time. Which command should the engineer use?

Accepted Answer

CREATE TABLE dev_db.public.fact_sales_clone CLONE prod_db.public.fact_sales;. Zero-copy cloning with CREATE TABLE ... CLONE creates a new table that references the same micro-partitions as the source, so no storage is duplicated at creation. CTAS would physically copy the data, CREATE TABLE ... LIKE would only copy the schema, and ALTER TABLE ... SET CLONE is not valid syntax. Only the CLONE form meets the no-extra-storage requirement.

Answer

CREATE TABLE dev_db.public.fact_sales_clone AS SELECT * FROM prod_db.public.fact_sales;

Answer

ALTER TABLE prod_db.public.fact_sales SET CLONE dev_db.public.fact_sales_clone;

Answer

CREATE TABLE dev_db.public.fact_sales_clone LIKE prod_db.public.fact_sales;