Develop a single, comprehensive script that handles all data transformations in one execution, minimizing the number of jobs submitted to the Azure Databricks cluster to reduce costs.

Design the data transformations as a series of smaller, modular functions that can be independently tested and reused, leveraging Apache Spark's distributed processing capabilities for efficiency and scalability.

Manually process the dataset row by row using a custom script to ensure precise control over each transformation step, despite the potential impact on processing time and scalability.

Utilize a mix of programming languages and tools for different transformation steps, selecting each based on its specific strengths for the task at hand, to optimize performance and flexibility.