DEA-C02 시험 - Snowflake실제시험문제와 답 - 91문항

Question No : 1

When using the `COPY INTO` command to load Parquet files in Snowflake, which of the following `FILE_FORMAT` options can directly influence the data loading process? (Choose 2.)

A.`DATE_FORMAT = 'YYYY-MM-DD'`
B.`COMPRESSION = 'GZIP'`
C.`TYPE = 'PARQUET'`
D.`STRIP_OUTER_ARRAY = TRUE`

정답:

Question No : 2

An analytics engineer runs `UNDROP TABLE my_table;` one hour after it was accidentally dropped. The command is successful.
What feature made this recovery possible?

A.Automatic Clustering
B.Zero-Copy Cloning
C.Fail-safe
D.Time Travel

정답:

Question No : 3

A data scientist has written a Python UDF to process data using a third-party library that is not included in the default Anaconda package channel provided by Snowflake.
What are the necessary steps the engineer must take to make this UDF work in Snowflake? (Choose 2.)

A.Upload the third-party library's Wheel or Zip file to a Snowflake internal stage.
B.The UDF must be rewritten in Java, as Python UDFs do not support third-party libraries.
C.Contact Snowflake support to have the package added to the default Anaconda channel.
D.Use the `IMPORTS` clause in the `CREATE FUNCTION` statement, referencing the path to the library in the stage.
E.Use `pip install` directly within the Python UDF code.

정답:

Question No : 4

A data architect is choosing a file format for landing Zone-Read-Decimal data from a legacy system into an external S3 stage. The downstream analytics in Snowflake require schema evolution (the ability to add or remove columns easily) and will perform aggregations on numeric fields.
Which file format would be the most suitable for this use case, considering both storage efficiency and compatibility with Snowflake's features?

A.XML
B.Avro
C.CSV
D.JSON

정답:

Question No : 5

An analytics engineer is responsible for a dashboard that runs the same complex aggregation query every hour on a very large, but slowly changing, base table. The query is slow and consumes significant warehouse credits. The business can tolerate the data in the dashboard being up to an hour stale.
To improve performance and reduce cost, the engineer decides to create a materialized view.
How does a materialized view address this problem?

A.It acts as a query cache, storing the result of the specific query text for faster retrieval.
B.It pre-computes and stores the result of the aggregation, so queries against the view read the stored result instead of querying the base table.
C.It automatically resizes the warehouse to a larger size just before the query is run.
D.It creates a zero-copy clone of the base table that is faster to query.

정답:

Question No : 6

A Snowflake Administrator creates a new role `SALES_ANALYST` but does not grant it to any other role or user.
Which role can manage (e.g., grant, revoke, drop) this new role?

A.Only the `SECURITYADMIN` role.
B.The role that created it (the owner) and any role higher in the hierarchy that inherits from the owner.
C.Only the `ACCOUNTADMIN` role.
D.Any role that has the `MANAGE GRANTS` privilege.

정답:

Question No : 7

A data engineer needs to implement a policy that prevents users in the `JUNIOR_ANALYST` role from seeing any rows in the `EMPLOYEES` table where the `SALARY` column is greater than 100,000. Users in the `SENIOR_ANALYST` role should be able to see all rows.
Which type of data governance policy is designed to achieve this row-level filtering?

A.Masking Policy
B.Session Policy
C.Row Access Policy
D.Password Policy

정답:

Question No : 8

An analytics engineer wants to improve the performance of queries that frequently filter a 10-billion-row table on the `USER_ID` column. The `USER_ID` column has very high cardinality.
Which feature is specifically designed to accelerate these types of highly selective, point-lookup queries?

A.Result Cache
B.Materialized Views
C.Search Optimization Service
D.Automatic Clustering

정답:

Question No : 9

A data architect is comparing data sharing methods.
What is a key benefit of using Snowflake Secure Data Sharing compared to traditional methods like sharing files via SFTP or S3?

A.The consumer is required to use the same cloud provider (e.g., AWS) as the data provider.
B.Data is shared live without being copied, meaning the consumer always sees the most up-to-date version and no additional storage costs are incurred by the consumer.
C.Data is physically copied to the consumer's account, providing them with an isolated copy.
D.It allows the consumer to grant write access back to the provider on the shared data.

정답:

Question No : 10

For a Snowpipe configured for automatic data ingestion, which of the following are necessary or recommended configurations? (Choose 2.)

A.The `ALTER PIP
B... REFRESH` command must be run manually on a schedule to trigger loading.
C.Event notifications must be configured on the cloud storage service (e.g., AWS S3) to inform Snowpipe of new files.
D.Snowpipe uses its own serverless compute resources for ingestion, separate from user-managed virtual warehouses.
E.Snowpipe can only be used to load CSV or JSON file formats.

정답:

Question No : 11

An analytics engineer is using the `COPY INTO` command to load a CSV file where text fields are enclosed in double quotes.
Which file format option should be used to correctly parse these fields?

A.`ENCLOSED_BY = '"'`
B.`FIELD_OPTIONALLY_ENCLOSED_BY = '"'`
C.`FIELD_DELIMITER = ','`
D.`ESCAPE_UNENCLOSED_FIELD = 'NONE'`

정답:

Question No : 12

What is the primary purpose of micro-partitions in the Snowflake architecture?

A.To enforce primary key constraints on tables.
B.To manage user concurrency on a virtual warehouse.
C.To serve as the unit of data storage and query pruning.
D.To store the results of recently executed queries.

정답:

Question No : 13

A data architect is designing a data governance framework where masking policies need to reference enterprise-wide metadata that is managed outside of Snowflake. For example, a column's data sensitivity level (`PII`, `CONFIDENTIAL`, `PUBLIC`) is stored in an external metadata catalog.
The requirement is to have a single masking policy that can be applied to any column. This policy must call an external API, passing the column name and table name as arguments, to fetch the sensitivity level. Based on the returned level, it will then apply the appropriate masking.
Which combination of Snowflake features is required to implement this? (Choose 2.)

A.Conditional Masking Policy Arguments
B.Secure Views
C.External Functions
D.Java UDFs
E.External Tables

정답:

Question No : 14

A data engineer is working with two large tables, `ORDERS` and `SHIPMENTS`, both containing a `REGION` column. A common query joins these two tables on their primary keys and filters for a specific region. To improve performance, the data engineer decides to create a multi-cluster warehouse.
How does a multi-cluster warehouse help optimize the performance of these queries, especially when multiple analysts are running them concurrently?

A.It automatically provisions additional, separate clusters to handle incoming queries in parallel, reducing queuing time.
B.It ensures that data for the same region from both tables is stored on the same micro-partition, speeding up the join.
C.It automatically rewrites the query to use a more efficient join algorithm.
D.It increases the amount of memory available for each individual query, preventing remote disk spilling.

정답:

Question No : 15

A data architect is designing an ingestion solution.
What is a primary difference between the Snowpipe REST API and the Snowpipe Streaming API?

A.The REST API uses Snowflake-managed compute, while the Streaming API requires the user to specify a running virtual warehouse.
B.The REST API is for batch loading files, while the Streaming API is for inserting individual rows without using files.
C.The REST API requires an Enterprise edition of Snowflake, while the Streaming API is available in all editions.
D.The REST API is lower latency than the Streaming AP

정답:

Snowflake DEA-C02 시험