This paper introduces Uniqueness Ratio as a pre-join metric for predicting privacy leakage in integrated databases. It analyzes how attribute combinations become identifying only after dataset joins, enabling early risk assessment before data integration using machine learning.