Create and run performance benchmark for the sampling cardinality estimation over index bounds. Compare two implementations:

• compute index keys from the sample and compute CE over them;
• transform index bounds to equivalent match expression and use 'matches' method.