Summary

This PR fixes a type inference issue in max_by and min_by aggregate functions when the first argument is a dictionary type.

Problem

Previously, when max_by or min_by received a dictionary as the first argument, the get_min_max_by_result_type function would only return the dictionary's value type, discarding information about other arguments. This could lead to type mismatches in query planning when the function result was used in subsequent operations.

Error during planning: Failed to coerce arguments to satisfy a call to 'max_by' function: coercion from [Dictionary(UInt32, Utf8), Int64] to the signature UserDefined failed No function matches the given name and argument types 'max_by(Dictionary(UInt32, Utf8), Int64)'

Solution

The fix modifies get_min_max_by_result_type to:

• Extract the dictionary's value type for the first argument (as before)
• Preserve all other argument types in the result vector
• Return the complete type information for proper query planning

Testing

• max_by with string/int, string/float, float/string, int/string combinations
• min_by with the same type combinations
• Both functions with dictionary-typed first arguments
• Validates correct result values and types

All tests pass successfully.