Free Snowflake SnowPro-Core Actual Exam Questions

A. Snowflake’s default for ACCESS HISTORY is definitely 7 days unless you pay for extended retention, so sticking with A makes the most sense here.

A. From what I know, default is 7 days for ACCESS HISTORY; anything longer needs special setup. So without extra info, A feels right.

Actually, I'd rule out A because GET_PRESIGNED_URL usually creates a time-limited access link, not just the file URL itself. That leaves C and B as closer options. B’s name suggests it gives an absolute path, but that might just be a file path, not a full hosted URL. C, BUILD_STAGE_FILE_URL, sounds like it specifically builds the hosted file URL from the stage and file path, which matches what the question asks. So I’d say it’s more likely C since it fits the idea of generating the URL directly.

Maybe B? GET_ABSOLUTE_PATH sounds like it would give you the full path from stage and file, which could be the Snowflake-hosted URL they’re asking for.

Probably A. Directory tables sound like they’re meant for organizing files or unstructured data given the name. The others (permanent, transient, temporary) seem more about how long the data sticks around rather than what kind of data they hold. So if the question is about supporting unstructured data specifically, directory tables might be the best fit since they could be designed to work with files or blobs directly. The rest are more general-purpose table types that can contain unstructured data but aren’t necessarily specialized for it.

It’s C. Temporary tables can store unstructured data just like permanent ones, but they’re session-scoped and disappear after use. So the key difference isn’t data type support, but lifetime.

D imo. The name ALLOW_CLIENT_MFA_CACHING suggests a longer cache time to reduce repeated prompts, so 8 hours fits better than shorter intervals like 2 or 4 hours.

C, since 4 hours is often the default cache time in similar MFA settings.

D/B? Time Travel lets you actively work with historical data, but Fail-safe is more like a backup after Time Travel expires. For ongoing maintenance, D seems more practical.

B imo, Fail-safe keeps your data longer for recovery after Time Travel ends.

A imo, VPrivate Snowflake restricts sharing so much it needs support help.

A/C? Virtual Private Snowflake definitely has the tightest controls, so needing support for sharing fits. Enterprise usually lets you do more self-service sharing without support. So if the question points to needing Snowflake Support, A feels more right than C. Standard is too basic but not necessarily blocking sharing outright, just limited features.

Yeah, C fits since materialized views do support joins across multiple tables.

Maybe C, since Snowflake’s docs say materialized views support joins, but not all join types.

D, because it explicitly says “all errors,” which suggests including previous partial loads too.

Maybe C could work since RETURN_ERRORS sounds like it deals with all errors detected during the load, which might include partially loaded files, unlike the others that focus on row counts.

Maybe C? It looks like a GET call for tiles, which might be unstructured data chunks like images or map pieces, unlike the insert or report options that sound more structured.

It’s D for me. The name “loadHistoryScan” sounds like it could handle raw scan data, which is often unstructured, unlike the others that seem to focus on reports or file uploads.

A imo, because Snowflake typically ties a masking policy to one data type only, but you can definitely set up multiple conditions and functions within that. B’s too limiting with just one condition and function.

I recall that Snowflake masking policies are designed around a single data type but can include multiple conditions and masking functions to handle different scenarios. So that rules out both C and D since they mention multiple data types. Between A and B, B seems too restrictive by allowing only one condition and function. So I’m going with A because it’s flexible enough for real-world use cases with more than one condition.

C/D? Since UNDROP only works if Time Travel is active, and the question doesn’t say so, I’d go with D as the default. B’s out because cloning isn’t restoring, it’s just copying. A doesn’t restore anything by itself, just enables a feature. So it really boils down to whether Time Travel is on or not, and without that detail, recreating (D) seems like the only guaranteed fix.

It’s C if Time Travel is enabled because UNDROP directly restores the dropped stage, which is cleaner than recreating it. If Time Travel isn’t on, then D makes sense since you have no rollback option. So without that info, I’d say C is the more precise restoration command when possible, not just recreating from scratch.

OWNERSHIP has to be given and basically can’t be revoked easily, so C.

Option C, OWNERSHIP generally sticks and can’t be revoked easily.

It’s C since Snowflake’s Query Profile focuses on detailed compute resource stats per query.

It’s C because the profile digs into resource-heavy queries, showing where compute is really being spent. B seems more about visuals, but this goes deeper into resource details.

I’m thinking A could be tossed out since queries without filters or aggregation usually run fast anyway. D’s off because search optimization is a different feature altogether. Between B and C, query acceleration likely helps more with big scans and filtering rather than high cardinality groups. So I’d go with B here too.

It’s B, since acceleration speeds up queries needing big data scans with filters.

Maybe D fits here too because transient tables are used for temporary data but with some persistence, and they might store metadata related to files staged for loading. Directory tables do track file metadata, but I wonder if transient tables hold file-level info when data is staged before final loading. Since external tables are about querying external data and temporary tables don’t persist beyond the session, D could make sense if the question targets file metadata stored temporarily but beyond session scope. Not totally sure though, but it’s worth considering given how transient tables man

It’s A. Directory tables are designed to track file-level metadata specifically for files in stages, which fits the question well. External tables mainly deal with querying external data but don’t store this kind of metadata internally. Temporary and transient tables are about table lifespan, not file metadata. So, if the focus is on metadata storage per file in a stage, directory tables make more sense here.