Which LINQ operators stream results and which ones buffer, and why does the difference matter?

In production I inspect the operator chain as one execution plan. Placing a selective Where or bounded Take before a buffering operation can reduce work when semantics permit it, while moving operators blindly can change results. For database-backed IQueryable, the provider may translate the chain into a server plan, so in-memory LINQ behavior is not the entire story. Once execution crosses into IEnumerable, the application owns enumeration and buffering. I test with realistic sizes, preserve cancellation at remote or async boundaries, and prefer explicit loops when the state machine or allocation profile needs to be obvious. A fluent expression is useful only while its execution cost remains understandable.