Collection and LINQ costs

Linear lookup inside a loop

foreach (var id in requestedIds)
{
    if (existingIds.Contains(id))
        Process(id);
}

When existingIds resolves to List<T>, the analyzer knows Contains is linear and can report BIG1001. With independent input sizes, the composed estimate preserves both variables rather than silently treating them as equal.

Materialization inside a loop

foreach (var group in groups)
{
    var active = group.Where(IsActive).ToList();
    Consume(active);
}

The terminal ToList enumerates and allocates inside the outer iteration. Supported symbol resolution can report BIG1002. Moving materialization outside the loop is correct only when the data and semantics allow reuse.

Ordering inside a loop

foreach (var batch in batches)
{
    var first = batch.OrderBy(x => x.Score).First();
    Consume(first);
}

Supported ordering inside iteration can report BIG1003. The right redesign may be pre-ordering, using a selection algorithm, or keeping the original code when input sizes are bounded and clarity wins.

Collection choice

Changing List<T> to HashSet<T> can replace linear lookup with expected O(1) lookup under hashing assumptions. It also changes ordering, duplicate behavior, memory use, and construction cost. The tool provides evidence; it does not choose the data structure for you.