Prepstellar

COF-C03 · Mock exam · 25 questions

COF-C03 mock exam: 25 free practice questions

Untimed here · timed and scored in the app

A free COF-C03 practice test with 25 exam-style questions drawn from the whole course and weighted by domain like the exam guide. Answer at your pace and read why each option is right or wrong.

  • 100Questions
  • 115 minTime
  • 750 (scaled 0–1,000)Passing score
  • Snowflake AI Data Cloud Features & Architecture 31%
  • Account Management and Data Governance 20%
  • Data Loading, Unloading, and Connectivity 18%
  • Performance Optimization, Querying, and Transformation 21%
  • Data Collaboration 10%

Exam details from the vendor’s exam guide, captured 2026-08-30.

Swipe, scroll or use ← →
  1. Q1 / 25

    Which table type provides row locking and enforced unique and referential integrity constraints?

    1. AApache Iceberg table
    2. BDynamic table
    3. CHybrid table
    4. DSnowflake table
    Show the answer

    Hybrid tables combine indexed random access, row locking, and integrity enforcement for transactional workloads.

    Next → 1 / 25
  2. Q2 / 25

    An unload must create date-based partition paths. Which option combination is valid?

    1. APARTITION BY with SINGLE set to TRUE
    2. BPARTITION BY with INCLUDE_QUERY_ID enabled
    3. CPARTITION BY with OVERWRITE set to TRUE
    4. DPARTITION BY with INCLUDE_QUERY_ID disabled
    Show the answer

    Partitioned output requires query IDs in filenames and disallows single-file output, overwrite, and disabling the query ID.

    Next → 2 / 25
  3. Q3 / 25

    A team wants to set the file size, compression, and internal organization of every Snowflake table manually. What should the architect explain?

    1. ASnowflake manages those storage details and table organization
    2. BEvery query must recreate those details before execution begins
    3. CEach virtual warehouse must configure those details independently
    4. DCloud services delegates those details to each client driver
    Show the answer

    Snowflake manages organization, file size, structure, compression, metadata, and statistics for data loaded into Snowflake tables.

    Next → 3 / 25
  4. Q4 / 25

    What happens to cached table data when its warehouse suspends?

    1. AThe cache moves into Query History
    2. BThe cache becomes a materialized view
    3. CThe cache persists for 24 hours
    4. DThe warehouse cache is dropped
    Show the answer

    Warehouse suspension discards the table-data cache; the separate persisted-result cache follows its own retention and reuse rules.

    Next → 4 / 25
  5. Q5 / 25

    Which three layers form the Snowflake architecture?

    1. ADatabase storage, compute, and cloud services
    2. BDatabase storage, networking, and client drivers
    3. CCompute, private hardware, and client software
    4. DCloud services, local storage, and user hardware
    Show the answer

    The architecture separates persisted data, workload execution, and platform coordination into database storage, compute, and cloud services.

    Next → 5 / 25
  6. Q6 / 25

    An unload must create date-based partition paths. Which option combination is valid?

    1. APARTITION BY with SINGLE set to TRUE
    2. BPARTITION BY with INCLUDE_QUERY_ID enabled
    3. CPARTITION BY with OVERWRITE set to TRUE
    4. DPARTITION BY with INCLUDE_QUERY_ID disabled
    Show the answer

    Partitioned output requires query IDs in filenames and disallows single-file output, overwrite, and disabling the query ID.

    Next → 6 / 25
  7. Q7 / 25

    A request is authenticated and parsed successfully, but no execution resources have yet run the SQL. Which handoff should occur next?

    1. AThe warehouse moves persisted data into private-cloud storage
    2. BDatabase storage authenticates the request on a compute node
    3. CCloud services dispatches the work to the selected warehouse
    4. DA client driver optimizes and executes the query without compute
    Show the answer

    Cloud services coordinates authentication, parsing, optimization, and dispatch; the selected warehouse then supplies execution compute.

    Next → 7 / 25
  8. Q8 / 25

    An unchanged staged file was already loaded successfully, but COPY must load it again. What is the relevant risk and option?

    1. APATTERN reloads it and automatically removes duplicates
    2. BPURGE reloads it and preserves one copy of every row
    3. CVALIDATION_MODE reloads it after checking all rows
    4. DFORCE reloads it and can duplicate table data
    Show the answer

    Normal COPY skips unchanged files with recorded load history; FORCE overrides that protection and can insert the data again.

    Next → 8 / 25
  9. Q9 / 25

    Which statement describes provider control over an existing share?

    1. AThe provider can revoke access only after copying all data
    2. BThe consumer must approve every provider object update
    3. CThe provider can revoke the share or object access at any time
    4. DThe consumer owns the share after importing its database
    Show the answer

    The provider keeps control over the share and its objects, including revocation; importing a database does not transfer that control.

    Next → 9 / 25
  10. Q10 / 25

    What happens to a warehouse's table-data cache when the warehouse is suspended?

    1. AThe cache becomes the cloud-services metadata store
    2. BThe cache persists permanently without using compute resources
    3. CThe cache moves into every other running warehouse
    4. DThe cache is dropped and rebuilt after resumption
    Show the answer

    Suspension saves idle credits but discards the warehouse cache, which can make initial queries slower until activity rebuilds it.

    Next → 10 / 25
  11. Halfway, at your pace

    In the app the mock exam is timed and scored like the real thing.

  12. Q11 / 25

    An unchanged staged file was already loaded successfully, but COPY must load it again. What is the relevant risk and option?

    1. APATTERN reloads it and automatically removes duplicates
    2. BPURGE reloads it and preserves one copy of every row
    3. CVALIDATION_MODE reloads it after checking all rows
    4. DFORCE reloads it and can duplicate table data
    Show the answer

    Normal COPY skips unchanged files with recorded load history; FORCE overrides that protection and can insert the data again.

    Next → 11 / 25
  13. Q12 / 25

    Which loading method writes rows directly to tables without staging files?

    1. APUT to a named stage
    2. BBulk COPY
    3. CSnowpipe Streaming
    4. DFile-based Snowpipe
    Show the answer

    Snowpipe Streaming removes the staged-file requirement by sending rows directly, unlike the two file-oriented COPY paths.

    Next → 12 / 25
  14. Q13 / 25

    Which description identifies structured data?

    1. ADocuments and images that have no inherent schema
    2. BNested JSON values that follow a flexible schema
    3. CRows and columns that follow a strict tabular schema
    4. DAudio files represented through the FILE data type
    Show the answer

    A strict tabular arrangement distinguishes structured data from flexible hierarchies and schema-free files.

    Next → 13 / 25
  15. Q14 / 25

    What does Snowflake's self-managed service model remove from the customer?

    1. ASelecting which persisted tables a query should read
    2. BChoosing a warehouse for statements that require compute
    3. CSelecting, installing, configuring, and managing service hardware
    4. DSubmitting SQL statements to process stored data
    Show the answer

    Snowflake manages the underlying hardware, software, updates, infrastructure, and ongoing tuning; users still direct their own workloads.

    Next → 14 / 25
  16. Q15 / 25

    Which loading method writes rows directly to tables without staging files?

    1. APUT to a named stage
    2. BBulk COPY
    3. CSnowpipe Streaming
    4. DFile-based Snowpipe
    Show the answer

    Snowpipe Streaming removes the staged-file requirement by sending rows directly, unlike the two file-oriented COPY paths.

    Next → 15 / 25
  17. Q16 / 25

    What is the default behavior of SINGLE for an unload?

    1. ASINGLE is FALSE, so multiple files are produced
    2. BSINGLE is NULL, so no output file is produced
    3. CSINGLE is TRUE, so exactly one file is produced
    4. DSINGLE is AUTO, so the stage chooses the count
    Show the answer

    Parallel multi-file output is the default; SINGLE must be set to TRUE when one output file is required.

    Next → 16 / 25
  18. Q17 / 25

    Which repeat execution satisfies a documented condition for full persisted-result reuse?

    1. AThe query text and contributing data remain unchanged
    2. BA table alias is added while the data remains unchanged
    3. CThe query text matches but contributing data has changed
    4. DKeyword capitalization changes while the data remains unchanged
    Show the answer

    Persisted-result reuse normally needs an exact textual match and unchanged contributing data; aliases and capitalization differences inhibit full reuse.

    Next → 17 / 25
  19. Q18 / 25

    Two teams query the same persisted tables, but one team's heavy workload must not reduce the other's compute performance. Which architectural choice addresses this requirement?

    1. AAssign each team an independent virtual warehouse
    2. BPlace both teams on the same warehouse compute cluster
    3. CCreate a separate cloud-services layer for each team
    4. DCopy every table into each warehouse's persistent storage
    Show the answer

    Independent virtual warehouses isolate compute performance while retaining access to centrally persisted data.

    Next → 18 / 25
  20. Q19 / 25

    What happens to cached table data when its warehouse suspends?

    1. AThe cache moves into Query History
    2. BThe cache becomes a materialized view
    3. CThe cache persists for 24 hours
    4. DThe warehouse cache is dropped
    Show the answer

    Warehouse suspension discards the table-data cache; the separate persisted-result cache follows its own retention and reuse rules.

    Next → 19 / 25
  21. Q20 / 25

    What is the default behavior of SINGLE for an unload?

    1. ASINGLE is FALSE, so multiple files are produced
    2. BSINGLE is NULL, so no output file is produced
    3. CSINGLE is TRUE, so exactly one file is produced
    4. DSINGLE is AUTO, so the stage chooses the count
    Show the answer

    Parallel multi-file output is the default; SINGLE must be set to TRUE when one output file is required.

    Next → 20 / 25
  22. Q21 / 25

    A team unloads to a Snowflake internal stage and then needs the files on a workstation. What is the correct sequence?

    1. ACOPY INTO <location>, then GET from the stage
    2. BGET from the table, then COPY INTO <table>
    3. CCOPY INTO <table>, then PUT to the workstation
    4. DPUT from the table, then query the stage
    Show the answer

    The unload command first creates files in the internal stage, and GET then downloads those staged files locally.

    Next → 21 / 25
  23. Q22 / 25

    What access do consumers receive to objects shared through Secure Data Sharing?

    1. ADelete access to the imported objects
    2. BOwnership access to the provider objects
    3. CWrite access to shared table data
    4. DRead-only access to the shared objects
    Show the answer

    Imported shared objects are read-only; the provider retains control while consumers can query through their imported database.

    Next → 22 / 25
  24. Q23 / 25

    Which queries can use new resources after a running warehouse is resized upward?

    1. AOnly queries that finished before the resize started
    2. BQueued and newly submitted queries after resources are provisioned
    3. CAll already-running queries from their original start time
    4. DOnly queries executed by other independent warehouses
    Show the answer

    A resize does not retroactively accelerate statements already executing; its provisioned capacity becomes available to queued and new work.

    Next → 23 / 25
  25. Q24 / 25

    Which statement describes provider control over an existing share?

    1. AThe provider can revoke access only after copying all data
    2. BThe consumer must approve every provider object update
    3. CThe provider can revoke the share or object access at any time
    4. DThe consumer owns the share after importing its database
    Show the answer

    The provider keeps control over the share and its objects, including revocation; importing a database does not transfer that control.

    Next → 24 / 25
  26. Q25 / 25

    A team unloads to a Snowflake internal stage and then needs the files on a workstation. What is the correct sequence?

    1. ACOPY INTO <location>, then GET from the stage
    2. BGET from the table, then COPY INTO <table>
    3. CCOPY INTO <table>, then PUT to the workstation
    4. DPUT from the table, then query the stage
    Show the answer

    The unload command first creates files in the internal stage, and GET then downloads those staged files locally.

    Next → 25 / 25
  27. That’s the whole mock exam

    Every question you miss comes back exactly when you’re about to forget it.

How to use this mock exam

Sit all 25 questions in one go: the mix covers every domain in the same proportion as the exam, so a low score points at the domain you skipped rather than at bad luck.

Read the explanation under every question, including the ones you got right — the reason an option is wrong is usually the thing being tested.

Then retake it in the app, where the mock exam is timed and scored and the questions you miss come back on a schedule.

The whole course, on your phone

Lessons you can read, audio you can listen to on the way to work, and practice that remembers what you got wrong.