Data scientists often work with large, evolving datasets that require efficient storage, processing, management, and analysis. Traditional file-based systems, where data is stored in separate files managed by individual applications, face several limitations when used for modern analytics and machine learning workflows. Although file systems can handle simple data storage requirements, they become inefficient and difficult to manage when dealing with large-scale, complex, and continuously changing datasets. One major limitation of traditional file-based systems is data redundancy and inconsistency. In a file-based approach, different applications often maintain their own copies of the same data. For example, a customer dataset may exist in multiple files used by different departments. When changes are made to one file, other copies may not be updated, leading to inconsistent and inaccurate data. For data scientists, inconsistent datasets can negatively affect data analysis