This paper introduces DREIFLUSS, an innovative, minimalist approach designed to tackle the Column Type Annotation (CTA) and Column Property Annotation (CPA) tasks in the SemTab challenge. DREIFLUSS efficiently employs semantic information from well-established knowledge graphs, DBpedia, and Schema.org, to improve the annotation process. Experimental evidence illustrates the superior performance of logistic regression models trained via DREIFLUSS, resulting in precise column-type annotations and insightful relationship predictions. The findings substantiate the significance of proper sampling technique while training a model, thereby boosting the accuracy and efficiency of table matching. This research illuminates a promising pathway to enhance table matching techniques, underlining the practical ramifications of DREIFLUSS for data integration and knowledge discovery endeavors.