SnowPro Core Certification (COF-C03)

Free SnowPro Core (COF-C03) Practice Questions

A handful of real practice questions from our SnowPro Core (COF-C03) bank — to give you a true feel for the style and difficulty before you sign up.

1. A team must pin client tools and SAML metadata to one Snowflake account identifier and wants the form that Snowflake recommends. Which identifier form is preferred?

   • A.The legacy account_locator.region.cloud form
   • B.The organization_name-account_name form used in the account host URL✓ Correct
   • C.A PrivateLink-only identifier returned by SYSTEM$WHITELIST_PRIVATELINK
   • D.A Marketplace listing identifier scoped to the provider organization

   Why: The organization_name-account_name form is the recommended identifier because it stays valid when Snowflake renames or relocates the account, which keeps client connection strings and SAML metadata working. The most tempting distractor is the legacy account_locator.region.cloud form, which is still valid but embeds the region and cloud and breaks SAML metadata if the account moves. SYSTEM$WHITELIST_PRIVATELINK returns firewall hostnames rather than account identifiers, and Marketplace listings use their own listing identifier separate from account addressing.

2. Unloading a 9 GB join result to an internal stage as Parquet should produce several files in parallel so a downstream Spark cluster can read them concurrently. Which COPY INTO clause keeps the unload split across multiple output files?

   • A.SINGLE = TRUE
   • B.MAX_FILE_SIZE = 5000000000
   • C.SINGLE = FALSE✓ Correct
   • D.INCLUDE_QUERY_ID = TRUE

   Why: Unload throughput depends on parallelism: SINGLE = FALSE is the default and lets the COPY split output across many files so warehouse threads write concurrently and downstream readers fan out the scan. SINGLE = TRUE forces a single output file and disables parallelism. MAX_FILE_SIZE caps each individual file but does not by itself create multiple files; the SINGLE clause is what gates the split. INCLUDE_QUERY_ID changes filenames only, leaving file count untouched.

3. Installing the Snowpark Python library on a local machine requires which pip package?

   • A.snowflake-snowpark-python✓ Correct
   • B.snowpark-python
   • C.pysnowflake-snowpark
   • D.snowflake-py-snowpark

   Why: Snowpark Python ships as the snowflake-snowpark-python package on PyPI, which provides the Session builder, DataFrame API, and UDF registration helpers. Installing any other name returns a package not found error or a different unrelated library. The closest distractor is snowpark-python because dropping the snowflake prefix matches common naming for shorter package names, but the published package keeps the snowflake prefix to disambiguate from third-party libraries. Always copy the name from the official documentation.

4. Which file-format option must be set to load a JSON file that contains a top-level array of objects as separate rows?

   • A.STRIP_OUTER_ARRAY = TRUE✓ Correct
   • B.STRIP_NULL_VALUES = TRUE
   • C.ENABLE_OCTAL = TRUE
   • D.DISABLE_AUTO_CONVERT = TRUE

   Why: JSON loads default STRIP_OUTER_ARRAY to FALSE, which causes a top-level array file to land as one VARIANT row containing the entire array. Setting STRIP_OUTER_ARRAY to TRUE removes the wrapping brackets so each element becomes its own row in the target table. STRIP_NULL_VALUES is the most tempting distractor because it sounds related to JSON parsing shape, but it removes null fields within each object and does not split an outer array into individual rows.

5. Nested customer event payloads arriving from a third-party webhook must land in a column that preserves the parsed document for downstream SQL navigation. Which Snowflake data type fits the requirement?

   • A.STRING that stores the raw text and defers parsing to query time
   • B.ARRAY that holds the top-level items in indexed order
   • C.BINARY that captures the compressed bytes from the source
   • D.VARIANT that stores the parsed semi-structured document for path access✓ Correct

   Why: The VARIANT type is built for semi-structured documents: it stores the parsed payload in a columnar internal layout and exposes path navigation through colon syntax and the FLATTEN function. STRING is the tempting alternative because the JSON arrives as text, yet storing the raw string forces every query to call PARSE_JSON before it can address a key. ARRAY restricts the document to an indexed list and BINARY discards SQL access entirely, both unsuitable for navigable storage.

SnowPro Core (COF-C03) Exam Details

Number of questions

100

Duration

115 minutes

Passing score

750 / 1000

Level

Associate

Delivered by

Snowflake

All figures should be confirmed on the official Snowflake page.

Frequently Asked Questions

Related certifications

• SnowPro DSA-C03
  Snowflake
• AWS SAA-C03
  Amazon Web Services (AWS)
• GCP - Associate Data Practitioner
  Google Cloud (GCP)
• AZ-104
  Microsoft Azure