Study SnowPro COF-C02 Pruning and Cost: key concepts, common traps, and exam decision cues.
On this page
This lesson matters because COF-C02 often rewards the answer that reduces unnecessary work rather than the answer that simply buys more compute. Pruning, selective access patterns, and cost guardrails are tightly connected.
Cost-and-tuning chooser
If the scenario is mostly about…
Strongest first focus
too much data scanned
pruning and query shape
runaway spend
warehouse control and workload design
repeated over-sizing
right-size first, then scale intentionally
What pruning is really solving
Problem
Better instinct
reading far more data than needed
improve filter selectivity and pruning-friendly access patterns
paying for idle or oversized compute
warehouse control and workload separation
mixing cost and performance into one vague complaint
classify scan waste vs compute waste first
Decision order that usually wins
Separate scan waste from compute waste.
If the query reads too much data, improve pruning before buying more compute.
If the warehouse is idle and still expensive, inspect suspend behavior and workload design.
Right-size repeatedly oversized warehouses instead of assuming bigger is safer.
Solve unnecessary work before solving speed with extra spend.
Common traps
Trap
Better rule
assuming bigger compute always fixes wide scans
poor pruning can still waste credits
ignoring warehouse suspend behavior
idle compute is a cost leak
reading every cost problem as a SQL-only problem
warehouse operations also matter
Scenario triage
Scenario clue
Stronger answer shape
“query touches too much data”
pruning and query-shape issue
“warehouse is idle but still burning credits”
suspend behavior and warehouse operations
“team keeps selecting larger warehouses for every complaint”