Backwards Scans

Backwards Scans

Backwards scans occur when SQL decides that it’s faster to start at the end of an index and scan towards the beginning. This happens frequently with the MAX() function, and when you order a query against the order of the index. If your index is ascending, and your query is ordered descending, for example. The optimizer chooses this rather than a potentially more expensive SORT later in the plan.

These typically aren’t the worst performance problems, but they should be looked at.

Backwards scans can’t be parallelized, and it may be easy to resolve by adding or adjusting an index.