Developed a flag-driven grouping separator for float and decimal formatting in the apache/datafusion repository, enhancing numeric output to match Java and Spark conventions. The solution introduced a new Rust-based formatting function, carefully integrated into the backend formatting pipeline without altering the public API or affecting existing users unless the feature flag is enabled. Unit tests were written to validate both the new logic and its seamless integration, ensuring numbers in scientific notation remain ungrouped. This work improves data readability in reporting and BI pipelines, reducing downstream formatting errors while maintaining backward compatibility and offering opt-in flexibility for users needing grouped output.