SnowPro Core (COF-C02) Practice Exam

What the SnowPro Core (COF-C02) exam covers

• Snowflake AI Data Cloud Features and Architecture172 questions
• Account Access and Security129 questions
• Performance Concepts108 questions
• Data Loading and Unloading72 questions
• Data Transformations144 questions
• Data Protection and Data Sharing93 questions

Free SnowPro Core (COF-C02) sample questions

A sample of 10 questions with answers and explanations. Sign up free to practice all 718.

1. Question 1Snowflake AI Data Cloud Features and Architecture

   What best describes Snowflake's overall architecture?

   • AA pure shared-nothing architecture where each node owns its own data and compute
   • BA pure shared-disk architecture where all nodes share the same compute cluster
   • CA multi-cluster, shared-data architecture that combines elements of shared-disk and shared-nothing designsCorrect
   • DA single-tier architecture where storage and compute run on the same physical server
   ✓ Correct answer: C

   Storage is centralized like a shared-disk system, while compute is split into independent MPP clusters like a shared-nothing system. This hybrid lets many compute clusters query the same data set without contention.

   Why the other options are wrong

   • ADescribes traditional shared-nothing systems, not how Snowflake separates storage from compute.
   • BDescribes shared-disk systems, but Snowflake does not require all compute to share one cluster.
   • DSnowflake explicitly separates storage from compute rather than co-locating them on one server.
2. Question 2Snowflake AI Data Cloud Features and Architecture

   A retail company only needs core Snowflake capabilities without advanced governance or multi-cluster scaling for concurrency, and is cost-sensitive. Which edition is most appropriate as a starting point?

   • AVirtual Private Snowflake (VPS)
   • BBusiness Critical
   • CStandardCorrect
   • DEnterprise with all add-ons enabled
   ✓ Correct answer: C

   It covers core functionality without the added cost of features this company does not need.

   Why the other options are wrong

   • AVPS is the most expensive, most isolated tier, overkill for basic needs.
   • BBusiness Critical targets highly regulated workloads with a higher price point.
   • DEnterprise offers more advanced features than needed here, at added cost.
3. Question 3Snowflake AI Data Cloud Features and Architecture

   Snowpipe's continuous, serverless ingestion is billed based on which of the following?

   • ASnowflake-managed compute resources used to process each file, billed per secondCorrect
   • BThe size of the virtual warehouse manually assigned to Snowpipe
   • CA flat monthly subscription fee regardless of usage
   • DThe number of SELECT statements run against the loaded table afterward
   ✓ Correct answer: A

   Snowpipe uses Snowflake-managed, serverless compute resources rather than a customer-specified virtual warehouse. Usage is billed per second based on the actual compute resources consumed while processing each incoming file.

   Why the other options are wrong

   • BSnowpipe uses Snowflake-managed serverless compute rather than a user-assigned warehouse, so warehouse size selection does not apply.
   • CSnowpipe billing is usage-based, per-second compute consumption, not a flat subscription fee.
   • DDownstream SELECT queries against the loaded table are billed separately based on whatever warehouse runs them.
4. Question 4Data Loading and Unloading

   Compared to a nightly scheduled bulk COPY INTO job, what latency characteristic does Snowpipe typically offer for newly arriving files?

   • AData typically becomes queryable within about a minute of arrivalCorrect
   • BData is only visible after exactly 24 hours
   • CSnowpipe has higher latency than any scheduled approach
   • DLatency is identical regardless of loading method
   ✓ Correct answer: A

   Because Snowpipe loads files as they arrive rather than waiting for a scheduled job, newly staged data commonly becomes queryable within about a minute, markedly faster than waiting for a nightly or hourly batch window.

   Why the other options are wrong

   • BThere is no fixed 24-hour delay in Snowpipe's near-real-time design.
   • CThis reverses the comparison; Snowpipe reduces latency versus scheduled batches.
   • DThe two approaches differ substantially in how soon data becomes available.
5. Question 5Performance ConceptsSelect all that apply

   Which TWO conditions best justify considering an explicit clustering key on a table? (Choose TWO)

   • AThe table is very large, multi-terabyte scale, and grows continuouslyCorrect
   • BQuery Profile shows poor pruning on commonly filtered columns despite the table's sizeCorrect
   • CThe table is small and rarely queried
   • DThe table only ever supports full table scans by design
   • EThe table has fewer than 100 micro-partitions total
   ✓ Correct answer: A, B

   Clustering keys pay off when scale and poor natural organization combine to hurt query performance in a way confirmed by Query Profile or clustering metadata. Small tables, tables designed only for full scans, or tables with very few partitions rarely benefit.

   Why the other options are wrong

   • CSmall, rarely queried tables typically see negligible benefit relative to the ongoing reclustering cost.
   • DIf a table only ever needs full scans, pruning provides no benefit regardless of clustering.
   • EA very small partition count means there is little to gain from reorganizing so few partitions.
6. Question 6Data Protection and Data Sharing

   Why does extending a table's Time Travel retention period from 1 day to 90 days generally increase storage cost?

   • ASnowflake must retain historical micro-partitions for changed or deleted rows across the full retention windowCorrect
   • BSnowflake duplicates the entire table into a separate physical database for the retention period
   • CLonger retention forces the table to be stored uncompressed for faster historical access
   • DLonger retention requires provisioning a dedicated virtual warehouse to hold the history
   ✓ Correct answer: A

   Time Travel storage cost scales with how much data has changed and how long it must be retained; a longer window means more historical versions of changed or deleted rows stick around, increasing overall storage consumption.

   Why the other options are wrong

   • BTime Travel does not duplicate the whole table into a new database; it retains the specific historical micro-partitions affected by changes.
   • CCompression is unrelated to retention length; historical data remains compressed like active data.
   • DWarehouses are compute resources; retained historical data is a storage concept, not tied to a dedicated warehouse.
7. Question 7Data Protection and Data Sharing

   What is a key cost consideration when enabling database replication to a secondary region?

   • AReplication cost replaces the primary database's storage cost rather than adding to it
   • BThe secondary consumes its own storage, and refreshes consume compute creditsCorrect
   • CReplication is entirely free and has no storage or compute cost implications
   • DReplication cost only applies once, at initial setup, and never recurs
   ✓ Correct answer: B

   Maintaining a secondary database means paying for its own storage footprint plus the compute credits consumed every time a REFRESH operation synchronizes it with the primary, on top of whatever the primary database already costs.

   Why the other options are wrong

   • AThe secondary's cost is additional, on top of the primary's own existing storage cost, not a substitute for it.
   • CReplication does have real cost implications, since the secondary database requires its own storage and refresh operations consume compute credits.
   • DCosts recur with every refresh operation and with ongoing storage of the secondary, not just a one-time setup charge.
8. Question 8Account Access and Security

   A network policy specifies both an allowed IP list and a blocked IP list. If an incoming connection's IP address matches an entry in the blocked list, what happens?

   • AThe connection is denied, even if that IP also appears in the allowed listCorrect
   • BThe connection is always permitted regardless of the allowed list
   • CThe blocked list is ignored unless MFA is also enabled
   • DThe entire network policy becomes disabled automatically
   ✓ Correct answer: A

   The blocked list acts as an explicit deny. If an IP address is present in both lists, Snowflake denies the connection because the block takes precedence, ensuring specifically excluded addresses cannot get through.

   Why the other options are wrong

   • BBeing on the blocked list results in denial, not automatic permission, regardless of the allowed list.
   • CThe blocked list is enforced independently of whether MFA is configured for the user.
   • DA blocked-list match does not disable the whole policy; it simply denies that particular connection.
9. Question 9Data TransformationsSelect all that apply

   Which TWO mechanisms can be used to convert a VARCHAR value into a DATE type in a Snowflake transformation query? (Choose TWO)

   • ATRY_CAST(col AS DATE)Correct
   • Bcol::DATECorrect
   • CFLATTEN(col)
   • DOBJECT_CONSTRUCT(col)
   • Ecol::ARRAY
   ✓ Correct answer: A, B

   Both TRY_CAST(col AS DATE) and the shorthand col::DATE convert a compatible string value into a DATE, with TRY_CAST returning NULL instead of erroring on an unparsable value.

   Why the other options are wrong

   • CFLATTEN explodes VARIANT arrays or objects into rows; it performs no type conversion.
   • DOBJECT_CONSTRUCT builds a VARIANT OBJECT from key-value pairs; it does not parse a string into a DATE.
   • Ecol::ARRAY would attempt to cast the value to an ARRAY type, not a DATE.
10. Question 10Data Transformations

    A stored procedure is written in LANGUAGE JAVASCRIPT and needs to dynamically build and run a SQL statement whose table name is determined at runtime. Which API does it use to run that SQL?

    • Asnowflake.execute()Correct
    • BTABLE()
    • CSYSTEM$STREAM_HAS_DATA()
    • DGET_DDL()
    ✓ Correct answer: A

    Inside a JavaScript stored procedure, the built-in snowflake object exposes an execute method that runs a SQL statement string constructed at runtime, such as one with a dynamically inserted table name. The other options are unrelated functions that do not execute arbitrary SQL from JavaScript.

    Why the other options are wrong

    • BTABLE() wraps a table function call in a SELECT statement; it does not execute dynamic SQL from a JS procedure.
    • CSYSTEM$STREAM_HAS_DATA checks whether a stream has unconsumed changes; it is unrelated to running dynamic SQL.
    • DGET_DDL returns the DDL text of an object; it does not execute a statement.

Who this SnowPro Core (COF-C02) practice exam is for

This practice set is for anyone preparing for the SnowPro Core (COF-C02) exam - from first-time candidates building a foundation to experienced Data practitioners doing a final review before test day. If you learn best by working through realistic questions and reading why each answer is right or wrong, it is built for you.

How to use this SnowPro Core (COF-C02) practice exam

1. Start with the free sample questions above to gauge your current baseline.
2. Read the full explanation on every question, including why each wrong option is wrong.
3. Track your weak domains and focus your study where you are losing the most marks.
4. Once you are scoring consistently well, take a timed, full-length mock exam.
5. Use your readiness score to decide when you are ready to book the real SnowPro Core (COF-C02) exam.

Related Data resources

• SnowPro Core (COF-C02) study guideKey concepts
• Data practice examsAll Data
• Certification pathWhere this fits
• Certification exam guides & tipsBlog
• Plans & pricingFree & paid
• Databricks Certified Data Engineer Associate practice examRelated

SnowPro Core (COF-C02) practice exam FAQ