Week 5 checkpoint. Ten questions on Domain 3: stages, file formats, COPY INTO loading and unloading, Snowpipe, Streams and Tasks, and the connectors and integrations. These are the five themes from Days 29 through 34. The mix mirrors the real COF-C03 pattern: mostly fact-recall, a couple of short scenarios, and three multi-select items where both […]
Week 4 checkpoint. Ten questions on Domain 2, the 20% of COF-C03 that covers governance, monitoring, and cost control. Real-exam style: shorter stems, mostly fact-recall, two scenarios. Multi-select count is three. Target 8 out of 10. Each wrong-answer explanation points to the day to re-read before Week 5. 01 Following Snowflake’s separation-of-duties best practice, which
Week 3 checkpoint. Ten questions covering Snowpark, Notebooks, Streamlit, Cortex AI, Cortex Search, Cortex Analyst, Snowflake ML, RBAC and DAC, system and secondary roles, and authentication. This is the first mixed-domain test in the series. The split is roughly 60% Domain 1 (AI/ML) and 40% Domain 2 (security). Stems are shorter than the Week 2
Week 2 checkpoint. Ten questions on Snowflake compute and storage, drawn from Days 8 through 13. The stems are tuned to the real COF-C03 mix: shorter wording, more fact-recall, only a couple of scenario questions. Target 8 out of 10. Every wrong-answer explanation points you to the exact day to re-read before Week 3. 01
Week 1 checkpoint. Ten questions covering architecture, editions, interfaces, the 12 database objects, and parameter precedence. These are the same five themes from Days 1 through 6. The mix mirrors the real COF-C03 pattern: a few straight recall items, two or three edition-boundary calls, plus at least one trickier question. On that one, every option
Day 38 added two serverless levers that left the warehouse alone. One offloaded a heavy scan. One built a search access path for point lookups. Today’s two tools change the data itself. A clustering key reorders rows so a query scans fewer micro-partitions. A materialized view stores a precomputed result so a repeated query skips
Day 39: Clustering Keys & Materialized Views for Snowflake Performance Read More »
Day 37 put each workload on its own warehouse. It also weighed scaling up against scaling out. Those levers change the warehouse itself. Today’s two levers leave the warehouse unchanged. They add serverless help on top of it. The Query Acceleration Service sends heavy scan work to a shared serverless pool. Search Optimization builds a
Day 38: Query Acceleration Service & Search Optimization in Snowflake Read More »
Day 36 taught you to read a single query’s profile. Today you zoom out to the warehouse that runs all of them. The job is to put the right work on the right warehouse, then read the columns that show when a warehouse is overloaded. One trap repeats across Domain 4: when concurrency is the
Day 37: Snowflake Workload Management & ACCOUNT_USAGE Performance Views Read More »
Week 6 opens Domain 4, the largest performance block on the COF-C03 (about 21% of the exam). Today is the diagnosis day. You learn to read the Query Profile and find the operator that costs the most. You also name the four problems it exposes: spilling, weak pruning, exploding joins, and queuing. Then you meet
Day 36: Reading the Snowflake Query Profile & Query Insights Read More »
No new material today. Five weeks in, you have Domain 3 covered: stages, file formats, COPY INTO loading and unloading, and Snowpipe. The week also covered Snowpipe Streaming, Streams and Tasks, and the drivers, connectors, and integrations that link Snowflake to the outside. Domain 3 is 18% of the exam. Today is a 30-minute closed-book
