Hello Popov,

Thank you for your detailed analysis and improvements! Let me explain the reasoning behind my approach and acknowledge where your version makes significant improvements:

1. About the "15" estimation:
You were correct - it was a rough estimation for the maximum expected digits before the decimal point in price values. Your approach of calculating exact sizes by finding maximum values first is much more precise and professional. It prevents any potential buffer overflows and is the proper enterprise-level solution.

2. Array Size Optimization:
Your implementation with GetCountOfDigits() and pre-calculation of exact array sizes is excellent. While my estimation approach might be slightly faster by avoiding the initial data pass, your solution provides guaranteed safety and precision, which is more important for production code.

3. FromDate Handling:
Your while-loop approach for calculating the FromDate bar index is more elegant than my solution. It provides better clarity and reliability.

4. Additional Improvements in Your Version:
- Adding the start date to the output is very useful for users
- Better error handling for bar count
- More comprehensive code structure and comments
- Consistent coding style throughout

Would you consider one potential optimization? For very large datasets, we could potentially cache the maximum values during the initial CopyRates to avoid the extra loop, though this would make the code more complex.

Your version strikes an excellent balance between performance, safety, and code quality. I appreciate how you improved upon my initial optimization attempt while maintaining production-grade reliability. I´ve removed my version to avoid confusion for new users, but can´t do so for the first post anymore (can´t edit it at this stage).

Thank you :-)