Smart farming increasingly depends on the integration of diverse data streams collected from IoT sensors, satellite imaging, and farm management systems. Although current Data Integration Frameworks (DIFs) provide effective support for data interoperability and system performance, most of them fall short in representing sustainability in a structured and actionable way. This study critically examines twelve DIF models drawn from agriculture, healthcare, transportation, and energy to evaluate how sustainability considerations are incorporated into their design. The findings show that while these frameworks are technically competent and emphasize scalability, they rarely include indicators such as carbon emissions, water efficiency, or energy use. A smaller group of models employs ontologies to improve semantic consistency across data, yet they still overlook sustainability as a foundational design objective. In response to these limitations, this paper introduces a Sustainability-Driven OntoDIF (SD-OntoDIF), a conceptual framework that integrates semantic technologies with sustainability metrics. By embedding ESG-aligned reasoning into data integration processes and utilizing tools such as artificial intelligence, edge computing, and ontological modelling, the proposed framework aims to support environmentally responsible and adaptive decision-making in smart farming.